Blood pressure and glycaemic effects of dapagliflozin versus placebo in patients with type 2 diabetes on combination antihypertensive therapy: a randomised, double-blind, placebo-controlled, phase 3 study

Articles

Blood pressure and glycaemic effects of dapagliflozin versus placebo in patients with type 2 diabetes on combination antihypertensive therapy: a randomised, double-blind, placebo-controlled, phase 3 study Michael A Weber, Traci A Mansfield, Valerie A Cain, Nayyar Iqbal, Shamik Parikh, Agata Ptaszynska

Summary Background Hypertension is a common comorbidity in patients with type 2 diabetes mellitus and a major risk factor for microvascular and macrovascular disease. Although the blood pressure-lowering effects of sodium–glucose cotransporter 2 (SGLT2) inhibitors are already established, guidance is needed on how to use these drugs in patients already receiving antihypertensive therapy. We aimed to compare blood pressure and glycaemic effects of the SGLT2 inhibitor dapagliflozin with placebo in patients with inadequately controlled type 2 diabetes mellitus and hypertension. Methods In this double-blind, placebo-controlled, phase 3 study we enrolled patients from 311 centres in 16 countries across five continents. Patients had uncontrolled type 2 diabetes (HbA1c 7·0%–10·5%; 53–91 mmol/mol) and hypertension (systolic 140–165 mm Hg and diastolic 85–105 mm Hg at both enrolment and randomisation, and a mean 24 h blood pressure of ≥130/80 mm Hg by ambulatory monitoring within 1 week of randomisation) and were receiving oral antihyperglycaemic drugs, insulin, or both, plus a renin–angiotensin system blocker and an additional antihypertensive drug. Using an interactive voice-response system, we randomly assigned (1:1) patients to dapagliflozin 10 mg once a day or to placebo, with randomisation stratified by additional antihypertensive drug use and insulin use at baseline, in a block size of two. The co-primary endpoints were changes in seated systolic blood pressure and HbA1c measured in the full analysis set, which included all patients who received at least one dose of study drug and had both a baseline and at least one post-baseline measurement of efficacy. This trial is registered with ClinicalTrials.gov, number NCT01195662. Findings Between Oct 29, 2010, and Oct 4, 2012, we randomly assigned 225 patients to dapagliflozin and 224 to placebo. Seated systolic blood pressure was significantly reduced in the group assigned to dapagliflozin (adjusted mean change from baseline –11·90 mm Hg [95% CI –13·97 to –9·82]) compared with those assigned to placebo (–7·62 mm Hg [–9·72 to –5·51]; placebo-adjusted difference for dapagliflozin −4·28 mm Hg [–6·54 to –2·02]; p=0·0002). Reductions in HbA1c concentrations were also significantly greater in patients assigned to dapagliflozin (adjusted mean change from baseline –0·63% [95% CI –0·76 to –0·50]) than in those assigned to placebo (–0·02% [–0·15 to 0·12]; placeboadjusted difference –0·61% [–0·76 to –0·46,]; p<0·0001). In a post-hoc analysis, we found difference in blood pressure versus placebo was greater in patients receiving a β blocker (−5·76 mm Hg [95% CI −10·28 to −1·23]) or a calciumchannel blocker (−5·13 mm Hg, [−9·47 to −0·79]) as their additional antihypertensive drug than in those receiving a thiazide diuretic (–2·38 mm Hg [–6·16 to 1·40]). Adverse events were similar in the dapagliflozin and placebo groups (98 [44%] patients vs 93 [42%], respectively, had at least one adverse event), with few adverse events related to renal function (1% vs <1%) or volume depletion (<1% vs 0%).

Lancet Diabetes Endocrinol 2015 Published Online November 24, 2015 http://dx.doi.org/10.1016/ S2213-8587(15)00417-9 This online publication has been corrected. The corrected version first appeared at thelancet.com/ diabetes-endocrinology on November 27, 2015 See Online/Comment http://dx.doi.org/10.1016/ S2213-8587(15)00457-X SUNY Downstate College of Medicine, Brooklyn, NY, USA (M A Weber MD); AstraZeneca, Fort Washington, PA, USA (T A Mansfield PhD); AstraZeneca, Wilmington, DE, USA (V A Cain MSc); AstraZeneca, Gaithersburg, MD, USA (N Iqbal MD, S Parikh MD); and Bristol-Myers Squibb, Princeton, NJ, USA (A Ptaszynska MD) Correspondence to: Dr Michael A Weber, SUNY Downstate College of Medicine, New York, NY 10021, USA [email protected]

Interpretation Dapagliflozin 10 mg significantly improved blood pressure and HbA1c and was tolerated similarly to placebo. Its blood pressure-lowering properties were particularly favourable in patients already receiving a β blocker or calcium-channel blocker. Dapagliflozin could benefit patients with type 2 diabetes who need a diuretic-like effect to optimise control of blood pressure, adding meaningful efficacy to antihypertensive drug regimens. Funding Bristol-Myers Squibb, AstraZeneca.

Introduction Hypertension is a common comorbidity that affects most patients with type 2 diabetes and contributes to the risk of cardiovascular disease and microvascular complications.1 Guidelines recommend that people with diabetes and hypertension be treated with an angiotensin-converting enzyme (ACE) inhibitor or angiotensin II receptor blocker and indicate that multidrug therapy is often needed to

achieve blood pressure goals (mostly defined as <140/90 mm Hg).1 Some glucose-lowering drugs are known to affect blood pressure—eg, glucagon-like peptide-1 (GLP-1) receptor agonists, thiazolidinediones, and (in some studies) dipeptidyl peptidase 4 (DPP-4) inhibitors produce small reductions in blood pressure.2 The most commonly prescribed classes of antihypertensive drugs are ACE inhibitors, angiotensin II

www.thelancet.com/diabetes-endocrinology Published online November 24, 2015 http://dx.doi.org/10.1016/S2213-8587(15)00417-9

1

Articles

Research in context Evidence before this study On Aug 11, 2015, we searched PubMed for articles with the search terms “SGLT2”, “T2DM”, and either “blood pressure” or “hypertension”, to identify reports of randomised controlled trials published in English with no date restrictions. We identified 143 articles, 23 of which were randomised trials. All studies reported reductions in blood pressure with dapagliflozin, canagliflozin, empagliflozin, ipragliflozin, remogliflozin, or sotagliflozin, with many studies reporting statistically significant decreases. Only three study protocols mandated that background drugs for blood pressure should be stable during the study treatment period. Most studies reported blood pressure as an efficacy endpoint (either as a co-primary endpoint [n=1], secondary endpoint [n=11], or exploratory endpoint [n=7]), although four trials included blood pressure as a safety outcome. For most trials, the primary endpoint was change from baseline in HbA₁C, with no other dapagliflozin studies and only one empagliflozin study (EMPA-REG BP) reporting change from baseline in blood pressure as a co-primary outcome. The EMPA-REG BP study, which was also the only other dedicated randomised trial in patients with type 2 diabetes and hypertension, showed significant reductions at week 12 in ambulatory 24 h systolic blood pressure (difference vs placebo −3·44 and −4·16 mm Hg with empagliflozin 10 mg and 25 mg, respectively; p<0·001 for both) and seated systolic blood pressure (difference vs placebo

receptor blockers, thiazide diuretics, calcium-channel blockers, and β adrenergic blocking agents (β blockers), each of which lowers blood pressure via distinct mechanisms.3 Sodium–glucose cotransporter 2 (SGLT2) inhibitors reduce glucose reabsorption in the proximal renal tubule, lowering blood glucose via an insulin-independent mechanism.4 Dapagliflozin is an orally-active, highlyselective SGLT2 inhibitor,5 which improves glycaemic control and is well tolerated, with a favourable safety profile in patients with type 2 diabetes.6–9 Dapagliflozin reduces blood pressure in patients with type 2 diabetes mellitus;10–13 an effect that has been attributed to osmotic diuresis, mild natriuresis, and weight loss.14 Furthermore, despite its diuretic effect, dapagliflozin decreases rather than increases serum uric acid12 and has not been associated with potassium abnormalities.15 Although the blood pressure-lowering effects of SGLT2 inhibitors are already established,10–13 guidance is needed on how to use these agents in patients already receiving the most commonly prescribed antihypertensive regimens; most typically comprising a renin-angiotensin system blocker plus either a diuretic, a calcium channel blocker, or a β blocker. In this phase 3 trial, we investigated the hypothesis that the blood pressurelowering and glucose-lowering effects of dapagliflozin 10 mg once a day would be superior to placebo in 2

−3·92 mm Hg and −4·80 mm Hg, respectively; p<0·001 for both), although no information was provided about what type or dose of antihypertensive drugs were taken by patients. Additionally, the results of a systematic review in which data were pooled from 27 randomised controlled trials of SGLT2 inhibitors (n=12 960) showed reductions in systolic blood pressure of 4·0 mm Hg (95% CI 3·5–4·4) relative to control. Added value of this study Dapagliflozin produced clinically meaningful reductions in blood pressure in patients with type 2 diabetes and hypertension, irrespective of the type of concomitant antihypertensive therapy, albeit with a smaller placebo-adjusted reduction in patients already receiving a thiazide diuretic. To the best of our knowledge, no previous study with an SGLT2 inhibitor has measured blood pressure reductions on the basis of underlying antihypertensive drugs. Implications of all the available evidence Our findings suggest a possible approach to identify patients with diabetes and hypertension most likely to benefit from blood pressure-lowering effects of SGLT2 inhibitor therapy. Dapagliflozin might be beneficial in patients with type 2 diabetes who require additional control of blood pressure. Further studies are needed to elucidate long-term outcomes with dapagliflozin in patients with type 2 diabetes and hypertension.

combination regimens of antihyperglycaemic and antihypertensive drugs in patients with inadequately controlled type 2 diabetes mellitus and hypertension.

Methods Study design and participants In this multicentre, randomised, double-blind, placebocontrolled, phase 3 trial, we enrolled patients at 311 centres in Australia, Canada, Colombia, Czech Republic, Denmark, Finland, Germany, Hungary, India, Ireland, Mexico, Poland, Puerto Rico, Romania, UK, and the USA. Eligible patients were those with type 2 diabetes, inadequate glycaemic control (HbA1c 7·0%–10·5%; 53–91 mmol/mol), and inadequately controlled hypertension (systolic 140–165 mm Hg and diastolic 85–105 mm Hg at both enrolment and randomisation, and a mean 24 h blood pressure of ≥130/80 mm Hg by ambulatory monitoring within 1 week of randomisation).16,17 Participants were required to be on a stable dose of at least one oral antihyperglycaemic drug for 6 weeks or more (≥12 weeks if thiazolidinedione) or a stable daily dose of insulin (as monotherapy or in combination with another oral antihyperglycaemic drug) for 8 weeks before enrolment, plus a stable, effective, therapeutic dose of a reninangiotensin system blocker (ACE inhibitor or angiotensin II receptor blockers) and an additional antihypertensive drug for 4 weeks or longer. The types of additional

www.thelancet.com/diabetes-endocrinology Published online November 24, 2015 http://dx.doi.org/10.1016/S2213-8587(15)00417-9

Articles

antihypertensive drugs permitted during the study were pre-specified as a thiazide diuretic, a thiazide-like diuretic, a calcium-channel blocker, a β blocker, an α adrenergic blocker, or a central α adrenergic agonist. Key exclusion criteria included serum creatinine of 177 μmol/L or lower (unless the patient was taking metformin, in which case exclusionary limits were serum creatinine ≥133 μmol/L for men and ≥124 μmol/L for women); estimated creatinine clearance of less than 60 mL per min, and significant hepatic disease. Full inclusion and exclusion criteria are listed in the appendix. The study protocol was approved by the institutional review board or independent ethics committee at each site and all patients provided written informed consent.

Randomisation and masking Patients were randomly assigned (1:1) to receive dapagliflozin 10 mg once daily or matched placebo. Randomisation was stratified by type of additional antihypertensive medication use and insulin use (yes vs no) at baseline in block sizes of two. An exploratory treatment arm in which 133 patients were randomised to dapagliflozin 5 mg was also included in the study, and patients continued treatment for the full study period (data not presented here). However, during the course of the trial, while the study was still blinded, a protocol amendment was initiated stipulating that patients would no longer be enrolled to the 5 mg arm. Randomisation was done by interactive voice response system. Randomisation codes were kept centrally at a BristolMyers Squibb facility in Lawrenceville, NJ, USA. Investigators and patients were masked to treatment allocation throughout the treatment period, after which the data were unmasked for reporting purposes. Masking was done through the identical appearance of the tablets, pill bottles, and labels.

Procedures The trial consisted of a qualification period (14 days or less after enrolment), a lead-in period (4 weeks), a doubleblind treatment period (12 weeks), and a follow-up period (1 week). During the qualification period, patients were given a baseline examination to determine eligibility. During the lead-in period, seated blood pressure and heart rate were measured 28 days, 14 days, and 1 day before the treatment period. Ambulatory blood pressure was assessed 1 day before treatment begun with Spacelabs model 90207 ambulatory blood pressure monitoring device (Spacelabs Healthcare, Snoqualmie, WA, USA). During the 12 week treatment period, patients were dispensed 10 mg of dapagliflozin or placebo tablets to be taken at home orally once a day in the morning. All antihyperglycaemic and antihypertensive drugs that participants were taking before enrolment were continued at the same dose throughout the study, with no new drugs permitted. Seated systolic blood pressure was measured at

each study visit (weeks 1, 2, 4, 8, and 12 after randomisation) between 0600 h and 1100 h after an initial resting period of 10 min. The mean was determined from three replicate measurements obtained at least 1 min apart. Patients had fasted for 8 h or longer and had last taken their antihypertensive and study drugs the previous day. Patients taking insulin who had an HbA1c value at enrolment of 7·0–7·4% had to reduce their daily insulin dose by 25% at randomisation and maintain this reduction throughout the treatment period to avoid hypoglycaemia and ensure patient safety. Any patients with a greater than 10% change in mean daily insulin dose after day 1 was discontinued. Rescue treatment for hyperglycaemia was not permitted. Any patient with inadequate glycaemic control (defined as a confirmed fasting plasma glucose >15·0 mmol/L) at any time after randomisation was discontinued. Open-label antihypertensive rescue treatment was available for severe (>180/110 mm Hg) or sustained (>165/105 and <180/110 mm Hg on two occasions) hypertension. Adverse events and laboratory values were assessed at every study visit, including at the follow-up visit 1 week after treatment. Adverse events were either volunteered by the patient or obtained through general questioning and examination at each study visit. Systolic blood pressure was also assessed for a final time at the end of the follow-up period (1 week after treatment with the study drug ended).

For the protocol see http:// astrazenecagrouptrials. pharmacm.com

Outcomes The co-primary endpoints were changes from baseline to week 12 in seated systolic blood pressure and HbA1c. Secondary endpoints included change from baseline to week 12 in 24 h ambulatory systolic blood pressure, seated diastolic blood pressure, and serum uric acid. Exploratory endpoints included the proportion of patients with improved blood pressure control (defined as <140/90 and <130/80 mm Hg); change from baseline to week 12 in ambulatory daytime, night-time, and trough systolic and diastolic blood pressures; and change from baseline to week 12 in fasting plasma glucose and bodyweight. To assess safety, we recorded the number of patients who had at least one adverse event, serious adverse event (defined as those events meeting the International Conference on Harmonisation Good Clinical Practice criteria7), and adverse event of special interest. The intensity of adverse events was classified according to the investigators’ judgment. We also assessed heart rate, measured orthostatic hypotension (defined as a decrease of >20 mm Hg in systolic blood pressure or >10 mm Hg in diastolic blood pressure from a supine to a standing position) and the proportion of patients receiving rescue treatment for severe or sustained hypertension, and took laboratory findings (serum concentrations of potassium, sodium, calcium, magnesium, phosphorous, chloride, bicarbonate and creatinine, estimated glomerular filtration rate [eGFR], haematocrit, aspartate aminotransferase, alanine aminotransferase, alkaline phosphatase, total

www.thelancet.com/diabetes-endocrinology Published online November 24, 2015 http://dx.doi.org/10.1016/S2213-8587(15)00417-9

3

Articles

2245 patients enrolled

1032 did not complete the enrolment period 934 did not meet eligibility criteria 65 withdrew consent 7 lost to follow-up 2 poor or no compliance 2 administrative reason by sponsor 21 other 1 adverse event

1213 completed the enrolment period

625 not randomised 497 did not meet eligibility criteria 69 withdrew consent 13 lost to follow-up 11 poor or no compliance 8 administrative reason by sponsor 6 adverse event 2 patient request to discontinue study treatment 19 other

588 randomised to treatment

139 not analysed 133 randomised to dapagliflozin 5 mg 6 no longer met eligibility criteria

224 randomly assigned to placebo

225 randomly assigned to dapagliflozin 10 mg

22 not completed 6 withdrew consent 3 lost to follow-up 2 administrative reason by sponsor 2 lack of efficacy 4 adverse events 3 other 1 request to discontinue study 1 no longer met study criteria

202 completed study

14 not completed 4 withdrew consent 5 no longer met study criteria 2 lost to follow-up 1 adverse event 1 administrative reason by sponsor 1 other*

211 completed study

Figure 1: Trial profile *Site disbanded before patient discontinuation information was obtained.

bilirubin, urine albumin to creatinine ratio [UACR]).

Statistical analysis This trial (MB102077) was originally planned as a four-arm, parallel-group trial to compare the addition of the following to patients regimens: dapagliflozin 2·5 mg, dapagliflozin 5 mg, dapagliflozin 10 mg, and placebo. The protocol specified that 1104 patients should be assigned into the four groups (276 in each) in a 1:1:1:1 ratio. On Aug 18, 2010, the protocol was amended to remove the 2·5 mg treatment arm. On Nov 1, 2011, enrolment to the 5 mg arm was discontinued when it became clear that the 10 mg dose 4

provided the best efficacy. On Aug 28, 2012, the protocol was also amended further to specify use a longitudinal repeated measures analysis instead of an ANCOVA model to account for missing data in the primary efficacy analysis. Due to these changes, 408 patients (204 in each group; assuming 5% do not have a post-baseline assessment) were needed for at least 80% power to detect a difference of 4 mm Hg in mean change from baseline in seated systolic blood pressure between the active and control group at a significance level of 0·05, assuming an SD of 14 mm Hg; and at least 94% power to detect a difference of 0·4% in mean change from baseline in HbA1c between the groups at significance level of 0·05, assuming an SD of 1·1%. Efficacy was assessed in the full analysis set, which included all patients who received at least one dose of study medication and had both a baseline and at least one post-baseline efficacy measurement. Data for patients after rescue for severe or sustained hypertension were excluded from blood pressure analyses from after the rescue date, but all data were included in analyses for change in HbA1c, serum uric acid, fasting plasma glucose, and bodyweight. The safety analysis set included all patients who received at least one dose of study drug, irrespective of rescue treatment. A hierarchical closed testing procedure was done across the primary and subsequent secondary endpoints to control family-wise type 1 error rate at a two-sided significance level of 0·05. For the primary efficacy analysis, we used a longitudinal repeated measures analysis using the direct likelihood method with fixed categorical effects of treatment, week, treatment-by-week interaction, and randomisation strata, and continuous fixed covariates of baseline seated systolic blood pressure value and baseline seated systolic blood pressure valueby-week interaction. If this primary analysis was statistically significant at the 0·05 level, the statistical test for the second co-primary endpoint (change from baseline in HbA1c) was done. This analysis used a similar longitudinal repeated measures analysis except that the continuous fixed covariates were baseline HbA1c and baseline HbA1c-by-week interaction. If the comparisons between the dapagliflozin 10 mg treatment group and the placebo group were significant at the 0·05 level for both co-primary endpoints, then the statistical tests for the secondary efficacy endpoints were done. All other efficacy outcomes were assessed by use of longitudinal repeated measures analyses, apart from 24 h ambulatory systolic blood pressure and daytime ambulatory systolic blood pressure, for which an ANCOVA model was used, with treatment group as an effect and baseline value and randomisation strata as covariates. All analyses were done with SAS/STAT version 8.2 or higher. This study is registered with ClinicalTrials.gov, number NCT01195662.

Role of the funding source The study funders were involved in the study design;

www.thelancet.com/diabetes-endocrinology Published online November 24, 2015 http://dx.doi.org/10.1016/S2213-8587(15)00417-9

Articles

Overall population

Additional antihypertensive subgroup Thiazide diuretic*

Calcium-channel blocker*

β blocker*

Placebo (n=61)

Placebo (n=59)

Placebo (n=224)

Dapagliflozin (n=225)

Placebo (n=77)

Dapagliflozin (n=92)

57·0 (51·0–62·0)

56·0 (50·0–62·0)

56·0 (51·0–61·0)

54·0 (48·5–61·0)

59·0 (51·0–62·0)

55·5 (49·0–60·0)

57·0 (51·0–62·0)

60·0 (52·0–64·0)

129 (58%)

118 (52%)

42 (55%)

50 (54%)

34 (56%)

30 (50%)

38 (64%)

30 (53%)

95 (42%)

107 (48%)

35 (46%)

42 (46%)

27 (44%)

30 (50%)

21 (36%)

27 (47%)

White

157 (70%)

160 (71%)

56 (73%)

68 (74%)

36 (59%)

33 (55%)

43 (72·9)

46 (81%)

Black

17 (8%)

19 (8%)

9 (12%)

7 (8%)

6 (10%)

9 (15%)

Asian

38 (17%)

34 (15%)

8 (10%)

11 (12%)

15 (27%)

15 (25%)

Age (years)

Dapagliflozin (n=60)

Dapagliflozin (n=57)

Sex Men Women Ethnic origin

Other Bodyweight (kg) Duration of type 2 diabetes (years) HbA₁C (%; mmol/mol) Fasting plasma glucose (mmol/L) Participants taking insulin

12 (5%)

12 (5%)

89·9 (18·4)

88·0 (20·5)

89·0 (17·7)

7·7 (5·9)

6·5 (4·6)

7·5 (6·5)

8·0% (1·0); 64 (11·3)

8·2% (1·0); 66 (10·7)

9·0 (2·4)

9·5 (2·5)

4 (5·2)

5 (5·4)

7·3 (5·0) 8·0% (1·0); 64 (10·6) 8·9 (2·4) 16 (7·1)

8·1% (0·9) 65 (10·1)

4 (5%)

9·0 (2·5) 18 (8·0)

6 (7%) 89·3 (22·4)

4 (7%) 87·8 (17·4)

3 (5%)

0 12 (20%) 4 (7%)

2 (4%) 6 (11%) 3 (5%)

87·5 (21·0)

92·0 (20·1)

87·8 (17·5)

7·0 (5·6)

7·0 (5·3)

8·1 (5·1)

8·1 (5·3)

7·8% (1·0); 62 (11·4)

8·1% (1·0); 65 (10·7)

8·2% (0·9); 66 (9·9)

8·0% (0·8); 64 (8·3)

8·3 (2·4)

8·4 (2·3)

9·4 (2·5)

8·9 (2·5)

3 (4·9)

7 (11·7)

7 (11·9)

4 (7·0)

Duration of hypertension (years)

9·3 (7·7)

9·4 (7·8)

8·7 (7·4)

8·5 (8·2)

9·1 (7·8)

8·5 (6·6)

9·3 (7·0)

10·6 (8·0)

Systolic blood pressure (mm Hg)

151·3 (6·7)

151·0 (7·9)

151·8 (6·9)

151·1 (8·6)

150·0 (6·4)

150·1 (7·4)

151·4 (6·9)

152·6 (7·3)

Diastolic blood pressure (mm Hg)

91·4 (4·8)

91·2 (4·8)

91·9 (5·0)

91·5 (4·7)

90·7 (4·8)

91·6 (4·9)

91·1 (4·3)

90·6 (4·8)

Selective vascular history 10 (5%)

12 (5%)

4 (5%)

3 (3%)

2 (3%)

1 (2%)

2 (3%)

7 (12%)

Stable angina

Coronary artery disease

4 (2%)

4 (2%)

2 (3%)

0

0

1 (2%)

2 (3%)

3 (5%)

Peripheral vascular disease

8 (4%)

1 (<1%)

3 (4%)

0

1 (2%)

0

2 (3%)

1 (2%)

Carotid artery disease

2 (1%)

2 (1%)

2 (3%)

0

0

1 (2%)

0

1 (2%)

Amputation History of dyslipidaemia

1 (<1%) 113 (50·4)

0 112 (49·8)

Recent cardiovascular event†

19 (9%)

21 (9%)

eGFR (mL/min per 1·73 m²)

87·0 (19·5)

84·8 (19·7)

1 (1%) 31 (40·3) 8 (10%) 86·7 (21·3)

0 37 (40%) 6 (7%) 85·1 (21·6)

0 34 (56%) 4 (7%) 84·4 (19·3)

0 32 (53%) 2 (3%) 86·9 (17·8)

0 32 (54%) 5 (9%) 90·9 (18·7)

0 33 (58%) 12 (21%) 83·1 (18·8)

Data are median (IQR), mean (SD), or n (%). ACE=angiotensin-converting enzyme. eGFR=estimated glomerular filtration rate. *Patients who did not take an additional antihypertensive drug from any of the three subgroups or who received drugs from more than one subgroup were excluded from subgroup analysis. †Previous myocardial infarction, percutaneous coronary intervention, coronary artery bypass graft, carotid endarterectomy or stenting, peripheral vascular surgery, cerebrovascular accident, transient ischaemic attack, congestive heart failure, or hospitalisation for unstable angina.

Table 1: Baseline demographics and disease characteristics, overall and by additional antihypertensive drug treatment

collection, analysis, and interpretation of data; and the decision to submit for publication. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

Results Between Oct 29, 2010, to Oct 4, 2012, we enrolled 1213 patients (figure 1). Of 449 patients in the full analysis set, 225 were randomly assigned to receive dapagliflozin 10 mg and 224 to receive placebo. Before protocol amendment, 133 patients were originally assigned to dapagliflozin 5 mg a day and continued this treatment for the full study period (data not shown). About 55% of participants received a stable dose of an ACE inhibitor (126 patients assigned to dapagliflozin vs 124 assigned to placebo) and about 45% received an angiotensin II receptor blocker (99 vs 99), in addition to a

thiazide diuretic, a thiazide-like diuretic, a calciumchannel blocker, a β blocker, or an α adrenergic blocker. No participants received a central α adrenergic agonist. Antihyperglycaemic drugs included metformin (203 [90%] of 225 assigned to dapagliflozin vs 206 [92%] of 224 assigned to placebo), sulfonylurea (105 [47%] vs 105 [47%]), DPP-4 inhibitors (16 [7%] vs 20 [9%]), thiazolidinedione (eight [4%] vs nine [4%]), insulin (18 [8%] vs 16 [7%]), acarbose (five [2%] vs three [1%]), and meglitinides (three [1%] vs none). Baseline demographics and disease characteristics were generally well balanced between treatment groups and additional antihypertensive subgroups (table 1), although history of cardiovascular diseases differed slightly between additional antihypertensive drug subgroups. Most patients had normal renal function or mild renal impairment, and less than a tenth of patients had an eGFR less than 60 mL/min per 1·73 m² at baseline.

www.thelancet.com/diabetes-endocrinology Published online November 24, 2015 http://dx.doi.org/10.1016/S2213-8587(15)00417-9

5

Articles

A

B

Seated systolic blood pressure Placebo Dapagliflozin

–2 –4 –6 –8 –10 –12

Adjusted mean change from baseline (%)

Adjusted mean change from baseline (mm Hg)

0

Baseline

2

219 224

218 221

4 Time (weeks) 213 220

8

12

–0·5 –0·6

205 212

199 205

Baseline Number at risk Placebo Dapagliflozin

217 220

D

Ambulatory systolic blood pressure*

0

20

–2

10

Adjusted mean change from baseline (μmol/L)

Adjusted mean change from baseline (mm Hg)

–0·3 –0·4

–0·8

Number at risk Placebo Dapagliflozin

–4 –6 –8 –10 –12 –14

4 Time (weeks) 214 219

8

12

207 211

197 204

8

12

207 212

198 204

Serum uric acid

0 –10 –20 –30 –40 –50

–16 0 Number at risk Placebo Dapagliflozin

–0·1 –0·2

–0·7

–14

C

HbA1C

0·2 0·1 0

2

4

6 8 10 12 14 16 18 20 22 Time since recording initiated (h)

24

176 178 181 184 184 184 182 183 184 184 185 151 175 181 182 184 184 185 185 184 185 185 183 158

Baseline Number at risk Placebo Dapagliflozin

217 220

4 Time (weeks) 214 219

Figure 2: Change over 12 weeks in seated systolic blood pressure (A), HbA₁C (B), 24 h ambulatory systolic blood pressure (C), and serum uric acid (D) in the full analysis set Error bars show 95% CIs. Blood pressure data excluded data after antihypertensive rescue treatment; data for HbA₁C and serum uric acid included data after antihypertensive rescue treatment. *24 h ambulatory monitoring was initiated between 0600 h and 1100 h for 24 h during week 12.

Patients assigned to dapagliflozin 10 mg had significantly greater reductions in mean seated systolic blood pressure from baseline to week 12 than did those assigned to placebo (adjusted mean change from baseline –11·90 mm Hg [95% CI –13·97 to –9·82] vs –7·62 mm Hg [–9·72 to –5·51], respectively; placeboadjusted difference −4·28 mm Hg [–6·54 to –2·02]; p=0·0002; figure 2A). At week 13, 1 week after the study drug was stopped, systolic blood pressure did not differ between the dapagliflozin and placebo groups (adjusted mean change from baseline –10·13 mm Hg [95% CI –12·26 to –8·01] vs –8·93 mm Hg [–11·08 to –6·78], respectively; placebo-adjusted difference −1·20 mm Hg [95% CI −3·56 to 1·15]). Reductions inHbA1c concentrations were significantly greater in patients assigned to dapagliflozin 10 mg than in those assigned to placebo (adjusted mean change from baseline –0·63% [95% CI –0·76 to –0·50] vs –0·02% [–0·15 to 0·12], respectively; placebo-adjusted difference –0·61% [–0·76 to –0·46], p<0·0001; figure 2B). Mean reductions from baseline values of 24 h ambulatory systolic blood pressure at week 12 were more 6

pronounced in patients in the dapagliflozin group than in those in the placebo group (placebo-adjusted difference −4·45 mm Hg [95% CI –7·14 to –1·76]; p=0·0012; figure 2C). At 12 weeks, dapagliflozin was associated with consistently greater reductions in ambulatory systolic blood pressure versus placebo at every timepoint assessed over a 24 h period. Mean seated diastolic blood pressure fell by 6·30 mm Hg in those assigned to dapagliflozin (95% CI 5·06 to 7·54) versus 5·33 mm Hg (4·08 to 6·59) in those assigned to placebo, respectively (placeboadjusted difference −0·97 mm Hg [95% CI –2·32 to 0·39]; p=0·16). The mean change from baseline to week 12 in fasting plasma glucose was −1·0 mmol/L (95% CI –1·4 to −0·6) in the dapagliflozin group compared with 0·2 mmol/L (–0·2 to 0·6) in the placebo group, with a placebo-adjusted difference of –1·2 mmol/L (–1·7 to –0·8). The change from baseline in serum uric acid at week 12 was –25·39 μmol/L (95% CI −35·71 to −15·07) in those assigned to dapagliflozin compared with –1·72 μmol/L (−12·12 to 8·69) in those assigned to placebo (placebo-adjusted difference −23·67 μmol/L [95% CI –33·70 to –13·64; figure 2D).

www.thelancet.com/diabetes-endocrinology Published online November 24, 2015 http://dx.doi.org/10.1016/S2213-8587(15)00417-9

Articles

A

B n

Baseline mean seated SBP, mm Hg (SD)

Difference vs placebo (95% CI)

Placebo

199

151·3 (6·8)

–4·28 (–6·54 to –2·02)

Dapagliflozin

205

151·0 (7·9)

Placebo

52

151·4 (6·9)

Dapagliflozin

51

152·6 (7·3)

Overall population

n

Baseline mean HbA1C, % (SD)

Difference vs placebo (95% CI)

Placebo

197

8·00 (0·96)

–0·61 (–0·76 to –0·46)

Dapagliflozin

204

8·09 (0·91)

Placebo

51

8·15 (0·92)

Dapagliflozin

51

8·01 (0·79)

Overall population

β blocker subgroup

β blocker subgroup –5·76 (–10·28 to –1·23)

Calcium-channel blocker subgroup

–0·80 (–1·10 to –0·50)

Calcium-channel blocker subgroup

Placebo

56

150·0 (6·4)

Dapagliflozin

58

150·1 (7·4)

Placebo

69

151·8 (6·9)

Dapagliflozin

80

151·1 (8·6)

–5·13 (–9·47 to –0·79)

Diuretic subgroup

Placebo

54

7·87 (1·02)

Dapagliflozin

58

8·13 (0·97)

Placebo

70

8·00 (1·01)

Dapagliflozin

79

8·18 (0·98)

–0·65 (–0·94 to –0·35)

Diuretic subgroup –2·38 (–6·16 to –1·40)

–20 –10 0 Adjusted mean change (mm Hg)

–0·45 (–0·70 to –0·19)

–1·0

C

–0·5 0 0·5 Adjusted mean change (%)

D n

Baseline mean ambulatory SBP, mm Hg (SD)*

Difference vs placebo (95% CI)

Placebo

186

149·2 (12·7)

–4·45 (–7·14 to 1·76)

Dapagliflozin

187

146·5 (10·4)

Placebo

46

149·8 (11·6)

Dapagliflozin

47

146·1 (11·1)

Overall population

n

Baseline mean serum uric acid, μmol/L (SD)

Difference vs placebo (95% CI)

Overall population

β blocker subgroup

Placebo

198 325·28 (78·92)

Dapagliflozin

204 334·95 (92·59)

–23·67 (–33·70 to –13·64)

β blocker subgroup –6·16 (–11·70 to –0·62)

Calcium-channel blocker subgroup

Placebo

51 313·46 (71·97)

Dapagliflozin

51 333·68 (88·03)

–36·88 (–56·51 to –17·25)

Calcium-channel blocker subgroup

Placebo

52

149·3 (13·3)

Dapagliflozin

50

146·3 (9·6)

Placebo

67

148·8 (13·8)

Dapagliflozin

75

146·9 (9·5)

–4·45 (–9·30 to 0·39)

Diuretic subgroup

Placebo

54 318·81 (75·53)

Dapagliflozin

58 332·49 (86·84)

–19·63 (–38·66 to –0·59)

Diuretic subgroup –4·92 (–9·42 to –0·43)

Placebo

70 336·06 (87·43)

Dapagliflozin

79 331·30 (100·52)

–20 –15 –10 –5 0 Adjusted mean change (mm Hg)

–13·09 (–29·15 to 2·97)

–60 –40 –20 0 20 40 Adjusted mean change (μmol/L)

Figure 3: Baseline and placebo-adjusted 12 week difference in seated SBP (A), HbA₁C (B), 24 h ambulatory SBP (C), and serum uric acid (D) by additional antihypertensive drug type, in the full analysis set Error bars show 95% CIs. Blood pressure data excluded data after antihypertensive rescue treatment; data for HbA₁C and serum uric acid included data after antihypertensive rescue treatment. Numbers in antihypertensive subgroups are only those with baseline and week 12 values available.*24 h ambulatory monitoring was initiated between 0600 h and 1100 h for 24 h on day 0 (baseline) and week 12. SBP=systolic blood pressure.

We did post-hoc subgroup analyses to assess efficacy variables by the class of additional antihypertensive drug received by patients during the treatment period (table 1). These analyses included only data from patients in the thiazide diuretic, calcium-channel blocker, and β blocker subgroups because the numbers of patients in the other subgroups were too small to make meaningful conclusions (<10 patients in either treatment group); patients who did not take an additional antihypertensive drug from any of these three subgroups or who took drugs from more than one subgroup were excluded from this subgroup analysis. Seated systolic blood pressure was reduced with dapagliflozin versus placebo to a greater extent in the β blocker subgroup and the calcium-channel blocker subgroup than in the thiazide diuretic subgroup (figure 3A). This difference in systolic blood pressure between the antihypertensive subgroups was reduced

when measured with 24 h ambulatory blood pressure monitoring (figure 3C). Seated diastolic blood pressure showed no relevant differences between the antihypertensive subgroups. The proportion of patients with improved blood pressure control (<140/90 mm Hg) was 38% (85 of 224) assigned to dapagliflozin versus 33% (72 of 219) assigned to placebo. The proportion of patients with improved blood pressure control (<130/80 mm Hg) was 8% (18 of 224) assigned to dapagliflozin versus 5% (12 of 219) assigned to placebo. Change from baseline in ambulatory daytime, night-time and trough systolic and diastolic blood pressures are shown in the appendix. Slightly greater reductions in daytime ambulatory systolic blood pressure were noted in patients assigned to dapagliflozin versus placebo (appendix), with greater reductions in the β blocker subgroup than in the calciumchannel blocker or thiazide diuretic subgroups.

www.thelancet.com/diabetes-endocrinology Published online November 24, 2015 http://dx.doi.org/10.1016/S2213-8587(15)00417-9

7

Articles

Overall population

One or more adverse event

Additional antihypertensive subgroup Thiazide diuretic

Calcium-channel blocker

β blocker

Placebo (n=224)

Dapagliflozin (n=225)

Placebo (n=77)

Dapagliflozin (n=92)

Placebo (n=61)

Dapagliflozin (n=60)

Placebo (n=59)

Dapagliflozin (n=57)

93 (42%)

98 (44%)

27 (35%)

46 (50%)

26 (43%)

21 (35%)

27 (46%)

24 (42%)

7 (9%)

12 (21%)

Adverse events that occurred in at least 10% of participants in either treatment group 32 (14%)

44 (20%)

20 (22%)

13 (21%)

8 (13%)

7 (12%)

Adverse event leading to discontinuation

Infections and infestations

4 (2%)

1 (<1%)

0

1 (1%)

2 (3%)

0

1 (2%)

0

One or more serious adverse event

2 (1%)

6 (3%)

1 (1%)

2 (2%)

1 (2%)

0

0

2 (4%)

Hypoglycaemia*

6 (3%)

13 (6%)

1 (1%)

2 (2%)

1 (2%)

6 (10%)

4 (7%)

5 (9%)

Renal function†

1 (<1%)

3 (1%)

0

2 (2%)

0

0

0

1 (2%)

Volume depletion†‡

0

1 (<1%)

0

1 (1%)

0

0

0

0

Genital infection†

4 (2%)

6 (3%)

1 (1%)

4 (4%)

3 (5%)

0

0

1 (2%)

Urinary tract infection†

2 (1%)

4 (2%)

0

3 (3%)

1 (2%)

0

0

1 (2%)

Special interest categories

Includes data after antihypertensive rescue. No deaths or malignancies were reported in the study. *We noted no major episodes of hypoglycaemia, defined as symptomatic episodes requiring external assistance with a capillary or plasma glucose value of less than 3 mmol/L and prompt recovery after glucose or glucagon administration; minor episodes of hypoglycaemia were defined as any symptomatic or non-symptomatic episode with a capillary or plasma glucose measurement of less than 3·5 mmol/L that did not qualify as a major episode; other episodes of hypoglycaemia were defined as investigator-reported episodes suggestive of hypoglycaemia that did not meet these criteria. †Based on a predefined list of preferred terms from Medical Dictionary for Regulatory Activities (MedDRA) version 15.1. ‡Defined as hypotension, dehydration, or hypovolaemia.

Table 2: Summary of adverse events for the 12 week treatment period

The effect of dapagliflozin on HbA1c concentrations was marginally more pronounced in the β blocker and calcium-channel blocker subgroups than in the thiazide diuretic subgroup (figure 3B). There were no clinically relevant differences in mean change from baseline in mean fasting plasma glucose between the additional antihypertensive subgroups (placebo-adjusted difference for thiazide diuretics –1·1 mmol/L [95% CI –1·9 to –0·4]; calcium-channel blocker –1·4 mmol/L [–2·3 to –0·5]; and β blockers –0·9 mmol/L [–1·8 to 0·0]). Reductions in serum uric acid were greater in the β blocker subgroup than the calcium-channel blocker or thiazide diuretic subgroups (figure 3D). The mean change in bodyweight from baseline to week 12 was −1·44 kg (95% CI −1·95 to −0·92) in patients assigned to dapagliflozin versus −0·59 kg (95% CI −1·11 to −0·07) in those assigned to placebo (placebo-adjusted difference −0·85 kg [−1·39 to −0·31]). The timecourse for bodyweight reduction seemed similar to that of systolic blood pressure reduction (appendix). Bodyweight was reduced with dapagliflozin compared with placebo to a greater extent in the β blocker subgroup (placebo-adjusted difference −1·51 kg [95% CI −2·64 to −0·39]) than in the calcium-channel blocker subgroup (−0·67 kg [−1·77, 0·43]) or thiazide diuretic subgroup (−0·49 kg [−1·43 to 0·44]). The incidence of adverse events was similar between the dapagliflozin (98 [44%] of 225 in the safety set) and placebo (93 [42%] of 224) group, with few events leading to discontinuation (table 2). Adverse events were more common in those taking dapagliflozin (n=46; 50%) than in those taking placebo (n=37; 35%) in the thiazide diuretic subgroup (table 2), whereas events were less common in those taking dapagliflozin in the other subgroups. More 8

patients had infections with dapagliflozin than with placebo in the thiazide diuretic subgroup (namely asymptomatic bacteriuria, urinary tract infection, influenza, and nasopharyngitis) and the β blocker subgroup (namely asymptomatic bacteriuria and nasopharyngitis), whereas these were less common with dapagliflozin than with placebo in the calcium-channel blocker subgroup. Few serious adverse events occurred during the study, but more patients in the dapagliflozin group (n=6; 3%) than in the placebo group (two [1%]) had a serious adverse event. Only one serious adverse event (in the placebo group) led to discontinuation. Hypoglycaemia was uncommon in both groups, but affected slightly more patients assigned to dapagliflozin than those assigned to placebo (table 2). No cases of major hypoglycaemia were reported. The incidence of hypoglycaemia was higher in patients assigned to dapagliflozin than in those assigned to placebo in the calcium-channel blocker subgroup whereas it was lower in patients in the thiazide diuretic or β blocker subgroups, although this effect might be due to higher insulin use in the calcium-channel blocker subgroup (table 1). Renal function-related adverse events were low in both treatment groups and across all antihypertensive subgroups (table 2). All renal events were mild or moderate in intensity and were mostly due to small changes in creatinine concentration. Renal function in the dapagliflozin 10 mg group remained stable over 12 weeks, with no clinically meaningful change in mean eGFR or urinary albumin-tocreatinine ratio (appendix). Laboratory values did not differ between the dapagliflozin and placebo groups at week 12 (appendix) and mean potassium and sodium concentrations remained stable in both groups.

www.thelancet.com/diabetes-endocrinology Published online November 24, 2015 http://dx.doi.org/10.1016/S2213-8587(15)00417-9

Articles

At baseline, measured orthostatic hypotension was reported in two patients in each treatment group. At week 12, measured orthostatic hypotension was noted in seven (3%) patients in the dapagliflozin 10 mg group and four (2%) in the placebo group, but no patients reported it as an adverse event. Only two patients (both in the placebo group) received rescue medication for severe or sustained hypertension. Seated heart rate at 12 weeks did not meaningfully differ from that at baseline for either group (−1·4 beats per min [bpm] with dapagliflozin 10 mg [baseline 77·1] vs −0·5 bpm with placebo [baseline 77·0]).

Discussion In this study, the addition of dapagliflozin to treatment regimens for 12 weeks was associated with lowered blood pressure and improved glycaemic control in patients with type 2 diabetes and hypertension inadequately controlled with up to two antihyperglycaemic drugs, a reninangiotensin system blocker, and an additional antihypertensive drug. Despite the concurrent antihypertensive treatments, the observed blood pressure effects of dapagliflozin seem potentially clinically relevant. However, for patients already receiving a renin–angiotensin system blocker plus a thiazide diuretic, dapagliflozin produced a rather small reduction in placebo-adjusted systolic blood pressure compared with those receiving a reninangiotensin system blocker plus a β blocker or calciumchannel blocker. This finding is not wholly unexpected, since adding a drug with a predominantly diureticdependent antihypertensive action14 to patients already receiving a thiazide drug might be less likely to have a major additive effect. This explanation is supported by our finding that the weight loss produced by dapagliflozin in the diuretic subgroup was less than that in the β blocker and calcium-channel blocker subgroups. Still, we cannot exclude the possibility that the slight placebosubtracted effect of dapagliflozin in the diuretic group might have been affected by the chance occurrence of a relatively large placebo effect. Our results show clinically meaningful, placebo-adjusted, decreases in systolic blood pressure of 5·1 mm Hg and 5·8 mm Hg in patients treated with a β blocker or calciumchannel blocker, respectively, which are of a similar scale to that which might be anticipated with a conventional diuretic added to ongoing dual antihypertensive therapy. For example, in an authoritative trial in patients with stage 2 hypertension (>160/100 mm Hg) receiving a reninangiotensin system blocker plus a calcium-channel blocker, the addition of hydrochlorothiazide 25 mg decreased systolic blood pressure by 6·2 mm Hg and diastolic blood pressure by 3·3 mm Hg compared with a renin– angiotensin system blocker plus calcium-channel blocker alone,18 an effect only marginally greater than we recorded in the present study despite the fact that the baseline blood pressure in the stage 2 patients was clearly higher than in our study population.

Dapagliflozin also significantly reduced 24 h ambulatory systolic blood pressure compared with placebo, with no difference between the three antihypertensive subgroups at 12 weeks. Since ambulatory blood pressure is regarded as a highly reliable basis for assessing hypertension,19 these results further corroborate the clinical benefits of dapagliflozin for patients with hypertension and type 2 diabetes. Dapagliflozin was not associated with significant reductions in diastolic blood pressure. However, systolic blood pressure is regarded as a stronger indicator of cardiovascular risk than is diastolic blood pressure,20 which has led to an increased emphasis on systolic blood pressure control in recent guidelines.3,21 Blood pressure-lowering effects have been reported for dapagliflozin in other clinical trials10–12 and in trials of other SGLT2 inhibitors.22,23 Furthermore, we have previously reported that dapagliflozin treatment is complementary to the use of renin–angiotensin system inhibitors.24 The results of the present study expand further our understanding of dapagliflozin and suggest that reduction of blood pressure is maintained irrespective of the class of additional antihypertensive drug, with a slightly greater effect in patients already receiving a β blocker or calciumchannel blocker than in patients already on a thiazide diuretic. Moreover, this trial is the first with dapagliflozin in which blood pressure was measured as a co-primary endpoint in patients with inadequately controlled hypertension and type 2 diabetes, with no changes or additions of background antihypertensive drugs allowed. We also note that patients in this study had a fairly long duration of hypertension (mean 9·4 years) and were uncontrolled despite dual combination therapy, suggesting that dapagliflozin might provide beneficial blood pressure effects in this setting. Moreover, dapagliflozin 10 mg was generally well tolerated and, unlike conventional thiazide diuretics,25 did not cause hypokalaemia. The glucose-lowering actions of dapagliflozin seen in this study have also been well documented in previous trials.7–9 The effects of dapagliflozin were similar across the antihypertensive subgroups, although variations in the effects of placebo meant that the placebo-adjusted reductions in HbA1c were slightly less substantial in the thiazide diuretic subgroup than in the β blocker or calcium-channel blocker subgroups. Reductions in serum uric acid with dapagliflozin have been reported previously,12,24 but our results confirm decreases with dapagliflozin in all antihypertensive subgroups (including the thiazide diuretic subgroup, in which there might be some concern about diureticinduced increases in uric acid25). Dapagliflozin was also associated with small reductions in bodyweight and, although it is unclear whether this effect contributed to the reduction in blood pressure, the timecourse for blood pressure reduction followed that of bodyweight reduction. Weight loss was slightly more pronounced among patients in the β blocker subgroup, although this result could have been affected by low patient numbers.

www.thelancet.com/diabetes-endocrinology Published online November 24, 2015 http://dx.doi.org/10.1016/S2213-8587(15)00417-9

9

Articles

Data from previous studies suggest that β blockers might be associated with weight gain.26 Our study had some limitations. Because of the strict criteria for enrolment, we had to include many study sites to recruit enough patients. However, protocol compliance and data accuracy were ensured by rigorous monitoring. Another possible limitation, as indicated earlier, is that the attenuated antihypertensive efficacy of dapagliflozin in patients already receiving a thiazide diuretic might have resulted simply from a relatively large placebo effect, although a smaller reduction in bodyweight seemed to confirm a lesser diuretic action of dapagliflozin in these thiazide-treated patients. A further limitation of the study was that very few patients with moderate renal impairment were included because SGLT2 inhibitors have a lower efficacy in this population.6 Additionally, despite the useful reductions in blood pressure, our study was not designed to measure major cardiovascular endpoints. However, an outcomes trial with dapagliflozin is currently in progress (Dapagliflozin Effect on CardiovascuLAR Events [DECLARE TIMI-58]; NCT01730534). Furthermore, in the recent EMPA-REG OUTCOME trial in patients with type 2 diabetes at high risk of cardiovascular events, the SGLT2 inhibitor empagliflozin reduced the risk of major cardiovascular events compared with placebo.27 Contributors All authors contributed to the conception and design of the study, the acquisition and interpretation of the data, and the drafting and critical revision of this report. Declaration of interests MAW received research funding from Bristol-Myers Squibb, and has been a consultant to AstraZeneca, Novartis, and Forest Pharmaceuticals. TAM is an employee of AstraZeneca and a shareholder of AstraZeneca and Bristol-Myers Squibb. VAC, NI, and SP are employees and shareholders of AstraZeneca. AP is an employee and shareholder of Bristol-Myers Squibb. Acknowledgments This study was funded by Bristol-Myers Squibb and AstraZeneca. Editorial assistance was provided by Helen Brereton inScience Communications, Springer Healthcare, London, UK, funding support for which was provided by AstraZeneca. References 1 American Diabetes Association. Standards of medical care in diabetes. Diabetes Care 2015; 38 (suppl 1): S1–S94. 2 Muskiet MH, Tonneijck L, Smits MM, Kramer MH, Lambers Heerspink HJ, van Raalte DH. Pleiotropic effects of type 2 diabetes management strategies on renal risk factors. Lancet Diabetes Endocrinol 2015; 3: 367–81. 3 Weber MA, Schiffrin EL, White WB, et al. Clinical practice guidelines for the management of hypertension in the community: a statement by the American Society of Hypertension and the International Society of Hypertension. J Clin Hypertens 2014; 16: 14–26. 4 Bailey CJ, Day C. SGLT2 inhibitors: glucuretic treatment for type 2 diabetes. Br J Diabetes Vasc Dis 2010; 10: 193–99. 5 Meng W, Ellsworth BA, Nirschl AA, et al. Discovery of dapagliflozin: a potent, selective renal sodium-dependent glucose cotransporter 2 (SGLT2) inhibitor for the treatment of type 2 diabetes. J Med Chem 2008; 51: 1145–49. 6 FDA. Farxiga (dapagliflozin) tablets. Highlights of prescribing information. US Food and Drug Administration, 2014. http:// www1.astrazeneca-us.com/pi/pi_farxiga.pdf (accessed Oct 15, 2015).

10

7

8

9

10

11

12

13

14

15

16 17

18

19 20 21

22

23

24

25 26

27

Bailey CJ, Gross JL, Pieters A, Bastien A, List JF. Effect of dapagliflozin in patients with type 2 diabetes who have inadequate glycaemic control with metformin: a randomised, double-blind, placebo-controlled trial. Lancet 2010; 375: 2223–33. Strojek K, Yoon KH, Hruba V, Elze M, Langkilde AM, Parikh S. Effect of dapagliflozin in patients with type 2 diabetes who have inadequate glycaemic control with glimepiride: a randomized, 24-week, double-blind, placebo-controlled trial. Diabetes Obes Metab 2011; 13: 928–38. Wilding JP, Norwood P, T’Joen C, Bastien A, List JF, Fiedorek FT. A study of dapagliflozin in patients with type 2 diabetes receiving high doses of insulin plus insulin sensitizers: applicability of a novel insulin-independent treatment. Diabetes Care 2009; 32: 1656–62. Woo V, Langkilde AM, Sugg J, Parikh S. Blood pressure reduction with dapagliflozin in patients with type 2 diabetes and cardiovascular disease. Circulation 2013; 128 (suppl 22): A10606. Woo V, Langkilde AM, Parikh S. Dapagliflozin, a novel antihyperglycemic agent that promotes urinary glucose excretion, reduces systolic blood pressure in patients with type 2 diabetes mellitus. Circulation 2011; 124 (suppl 21): A9520 (abstr). Basile J, Ptaszynska A, Ying L, Sugg J, Parikh S. The effects of dapagliflozin on cardiovascular risk factors in patients with type 2 diabetes mellitus. Circ Cardiovasc Qual Outcomes 2012; 5: A59. Sjostrom CD, Johansson P, Ptaszynska A, List J, Johnsson E. Dapagliflozin lowers blood pressure in hypertensive and non-hypertensive patients with type 2 diabetes. Diab Vasc Dis Res 2015; 12: 352–58. Lambers Heerspink HJ, de Zeeuw D, Wie L, Leslie B, List J. Dapagliflozin a glucose-regulating drug with diuretic properties in subjects with type 2 diabetes. Diabetes Obes Metab 2013; 15: 853–62. Yavin Y, Mansfield TA, Ptaszynska A, et al. Hyperkalemia incidence with the SGLT2 inhibitor dapagliflozin. Diabetes 2014; 63 (suppl 1): 1086-P. Standards of medical care in diabetes—2009. Diabetes Care 2009; 32 (suppl 1): S13–61. Chobanian AV, Bakris GL, Black HR, et al. The seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: the JNC 7 report. JAMA 2003; 289: 2560–72. Calhoun DA, Lacourciere Y, Chiang YT, Glazer RD. Triple antihypertensive therapy with amlodipine, valsartan, and hydrochlorothiazide: a randomized clinical trial. Hypertension 2009; 54: 32–39. Pickering TG, Shimbo D, Haas D. Ambulatory blood-pressure monitoring. N Engl J Med 2006; 354: 2368–74. Kannel WB. Elevated systolic blood pressure as a cardiovascular risk factor. Am J Cardiol 2000; 85: 251–55. NICE. The clinical management of primary hypertension in adults. Clinical Guideline 127. UK National Institute for Clinical Excellence, 2011. https://www.nice.org.uk/guidance/cg127 (accessed Oct 15, 2015). Baker WL, Smyth LR, Riche DM, Bourret EM, Chamberlin KW, White WB. Effects of sodium-glucose co-transporter 2 inhibitors on blood pressure: a systematic review and meta-analysis. J Am Soc Hypertens 2014; 8: 262–75, e9. Tikkanen I, Narko K, Zeller C, et al. Empagliflozin reduces blood pressure in patients with type 2 diabetes and hypertension. Diabetes Care 2015; 38: 420–28. Weber MA, Mansfield TA, Alessi F, Ptaszynska A. Effects of dapagliflozin on blood pressure in diabetic patients with hypertension inadequately controlled by a renin-angiotensin system blocker. Circulation 2013; 128: A13144 (abstr). Sica DA, Carter B, Cushman W, Hamm L. Thiazide and loop diuretics. J Clin Hypertens 2011; 13: 639–43. Sharma AM, Pischon T, Hardt S, Kunz I, Luft FC. Hypothesis: β-adrenergic receptor blockers and weight gain: a systematic analysis. Hypertension 2001; 37: 250–54. Zinman B, Wanner C, Lachin JM, et al. Empagliflozin, cardiovascular outcomes, and mortality in type 2 diabetes. N Engl J Med 2015; published online Sept 17. DOI:10.1056/NEJMoa1504720.

www.thelancet.com/diabetes-endocrinology Published online November 24, 2015 http://dx.doi.org/10.1016/S2213-8587(15)00417-9