0
Original Research |

Long-Term Efficacy of Dapagliflozin in Patients With Type 2 Diabetes Mellitus Receiving High Doses of Insulin: A Randomized Trial FREE

John P.H. Wilding, DM; Vincent Woo, MD; Norman G. Soler, MD, PhD; Andrea Pahor, MD; Jennifer Sugg, MS; Katja Rohwedder, MD; Shamik Parikh, MD, for the Dapagliflozin 006 Study Group
[+] Article and Author Information

Presented in part as an oral presentation at the 70th Scientific Sessions of the American Diabetes Association, Orlando, Florida, 25–29 June 2010, and as a poster presentation at the 46th Annual Meeting of the European Association for the Study of Diabetes, Stockholm, Sweden, 20–24 September 2010.

Grant Support: By AstraZeneca. Medical writing assistance for this manuscript was provided by Julian Martins, MA, MBBS, of inScience Communications, and this assistance was funded by Bristol-Myers Squibb and AstraZeneca.

Potential Conflicts of Interest: Disclosures can be viewed at www.acponline.org/authors/icmje/ConflictOfInterestForms.do?msNum=M11-0685.

Reproducible Research Statement:Study protocol, statistical code, and data set: Not available.

Requests for Single Reprints: John P.H. Wilding, DM, Diabetes and Endocrinology Research Group, Department of Obesity and Endocrinology, Clinical Sciences Center, University Hospital Aintree, Longmoor Lane, Liverpool L9 7AL, United Kingdom; e-mail, mailto:j.p.h.wilding@liverpool.ac.uk.

Current Author Addresses: Dr. Wilding: Diabetes and Endocrinology Research Group, Department of Obesity and Endocrinology, Clinical Sciences Center, University Hospital Aintree, Longmoor Lane, Liverpool L9 7AL, United Kingdom.

Dr. Woo: Section of Endocrinology, Health Sciences Center, University of Manitoba, Winnipeg, Manitoba R3T 2N2, Canada.

Dr. Soler: Springfield Diabetes and Endocrine Center, Hospital Sisters Health System Medical Group, 2501 Chatham Road, Suite 300, Springfield, IL 62704.

Drs. Pahor and Rohwedder: AstraZeneca, Tinsdaler Weg 183, D-22880 Wedel, Germany.

Ms. Sugg and Dr. Parikh: AstraZeneca, 1800 Concord Pike, Box 15437, Wilmington, DE 19850-5437.

Author Contributions: Conception and design: J.P.H. Wilding, A. Pahor, J. Sugg, S. Parikh.

Analysis and interpretation of the data: J.P.H. Wilding, V. Woo, J. Sugg, K. Rohwedder, S. Parikh.

Drafting of the article: A. Pahor, K. Rohwedder.

Critical revision of the article for important intellectual content: J.P.H. Wilding, V. Woo, N.G. Soler, S. Parikh.

Final approval of the article: J.P.H. Wilding, V. Woo, N.G. Soler, A. Pahor, J. Sugg, K. Rohwedder, S. Parikh.

Provision of study materials or patients: J.P.H. Wilding, N.G. Soler.

Statistical expertise: J. Sugg.

Collection and assembly of data: J.P.H. Wilding, A. Pahor.

*Appendix 1 lists the investigators of the Dapagliflozin 006 Study Group.


From University Hospital Aintree, Liverpool, United Kingdom; University of Manitoba, Winnipeg, Canada; Springfield Diabetes and Endocrine Center, Springfield, Illinois; and AstraZeneca, Wedel, Germany, and Wilmington, Delaware.


Ann Intern Med. 2012;156(6):405-415. doi:10.7326/0003-4819-156-6-201203200-00003
Text Size: A A A

Background: Dapagliflozin, a selective inhibitor of sodium–glucose cotransporter 2, may improve glycemic control with a lower dose of insulin and attenuate the associated weight gain in patients with inadequate control despite high doses of insulin.

Objective: To evaluate the efficacy and safety of adding dapagliflozin therapy in patients whose type 2 diabetes mellitus is inadequately controlled with insulin with or without oral antidiabetic drugs.

Design: A 24-week, randomized, placebo-controlled, multicenter trial followed by a 24-week extension period. An additional 56-week extension period is ongoing. (ClinicalTrials.gov registration number: NCT00673231)

Setting: 126 centers in Europe and North America from 30 April 2008 to 19 November 2009.

Patients: 808 patients with inadequately controlled type 2 diabetes mellitus receiving at least 30 U of insulin daily, with or without up to 2 oral antidiabetic drugs.

Intervention: Patients were randomly assigned in a 1:1:1:1 ratio and allocated with a computer-generated scheme to receive placebo or 2.5, 5, or 10 mg of dapagliflozin, once daily, for 48 weeks.

Measurements: The primary outcome was change in hemoglobin A1c from baseline to 24 weeks. Secondary outcomes included changes in body weight, insulin dose, and fasting plasma glucose level at 24 weeks and during the 24-week extension period. Adverse events were evaluated throughout both 24-week periods.

Results: 800 patients were analyzed. After 24 weeks, mean hemoglobin A1c decreased by 0.79% to 0.96% with dapagliflozin compared with 0.39% with placebo (mean difference, −0.40% [95% CI, −0.54% to −0.25%] in the 2.5-mg group, −0.49% [CI, −0.65% to −0.34%] in the 5-mg group, and −0.57% [CI, −0.72% to −0.42%] in the 10-mg group). Daily insulin dose decreased by 0.63 to 1.95 U with dapagliflozin and increased by 5.65 U with placebo (mean difference, −7.60 U [CI, −10.32 to −4.87 U] in the 2.5-mg group, −6.28 U [CI, −8.99 to −3.58 U] in the 5-mg group, and −6.82 U [CI, −9.56 to −4.09 U] in the 10-mg group). Body weight decreased by 0.92 to 1.61 kg with dapagliflozin and increased by 0.43 kg with placebo (mean differences, −1.35 kg [CI, −1.90 to −0.80 kg] in the 2.5-mg group, −1.42 kg [CI, −1.97 to −0.88 kg] in the 5-mg group, and −2.04 kg [CI, −2.59 to −1.48 kg] in the 10-mg group). These effects were maintained at 48 weeks. Compared with the placebo group, patients in the pooled dapagliflozin groups had a higher rate of hypoglycemic episodes (56.6% vs. 51.8%), events suggesting genital infection (9.0% vs. 2.5%), and events suggesting urinary tract infection (9.7% vs. 5.1%).

Limitation: Insulin doses were not titrated to target, and the study was not designed to evaluate long-term safety.

Conclusion: Dapagliflozin improves glycemic control, stabilizes insulin dosing, and reduces weight without increasing major hypoglycemic episodes in patients with inadequately controlled type 2 diabetes mellitus.

Primary Funding Source: AstraZeneca and Bristol-Myers Squibb.

Editors' Notes
Context

  • Persons with type 2 diabetes who receive ongoing insulin therapy may still have inadequate glycemic control, but increasing the insulin dose can result in troubling or dangerous side effects. Dapagliflozin inhibits the renal absorption of glucose and improves glycemic control as monotherapy or when added to metformin in clinical trials.

Contribution

  • In this 24-week randomized trial, adding dapagliflozin to insulin therapy for type 2 diabetics with inadequate glycemic control resulted in improved hemoglobin A1c values and weight loss compared with placebo.

Caution

  • Genital infections were more common with dapagliflozin, and the study was not designed to evaluate long-term safety.

Implication

  • Dapagliflozin may be helpful for hyperglycemia in type 2 diabetes that is poorly controlled despite insulin therapy.

—The Editors


Nearly 1 in 4 patients with type 2 diabetes mellitus eventually requires insulin therapy (1) because of progressive deterioration of glycemic control (2). However, patients with inadequately controlled diabetes despite substantial doses of insulin are particularly challenging to treat. Therapeutic options for these patients are limited, because increasing the dosage of insulin also increases the risks for weight gain, hypoglycemia, fluid retention, and congestive heart failure (38).

Sodium–glucose cotransporter 2 is a low-affinity, high-capacity transporter that mediates renal glucose reabsorption (9). Dapagliflozin, a competitive and highly selective inhibitor of sodium–glucose cotransporter 2 (1011), reduces renal glucose reabsorption, increases renal glucose excretion, and reduces hyperglycemia in a dose-dependent manner (12). Because dapagliflozin acts independently of insulin, it may provide additional glycemic control when used with insulin (3). Moreover, the caloric loss and osmotic diuresis secondary to increased urinary glucose excretion may counter insulin-related weight gain and fluid retention, respectively.

In clinical trials of up to 24 weeks in patients with type 2 diabetes, dapagliflozin improved glycemic control and body weight when administered as monotherapy (1315) or when added to therapy with metformin (16) or glimepiride (17). In a 12-week pilot study of 71 patients with type 2 diabetes mellitus whose initial insulin dosages were reduced by 50% (18), dapagliflozin reduced hemoglobin A1c (HbA1c), improved fasting and postprandial blood glucose levels, and lowered body weight compared with placebo.

The aim of our large, long-term study was to evaluate glycemic efficacy, weight loss, and safety over 48 weeks of dapagliflozin therapy in patients whose type 2 diabetes was inadequately controlled with insulin therapy.

Our double-blind, placebo-controlled, parallel-group trial was conducted at 126 centers worldwide from 30 April 2008 to 19 November 2009. The protocol was approved by institutional review boards and independent ethics committees, and all participants gave written, informed consent. The study was conducted in accordance with the principles of the Declaration of Helsinki (2004 version) and the requirements of good clinical practice.

After a 2-week enrollment, patients were randomly assigned on a 1:1:1:1 basis to receive placebo or 2.5, 5, or 10 mg of dapagliflozin, once daily, in addition to open-label therapy with their usual daily dose of insulin and existing oral antidiabetic drugs (OADs). Treatment continued for 24 weeks, during which patients, investigators, study monitors, and study sponsors were blinded. This period was followed by a 24-week extension period. A further 56-week extension period is in progress. Patients, investigators, and study monitors remained blinded during the extension periods.

A computer-generated, stratified, block-randomization schedule containing stratum, randomization code, and treatment was provided by AstraZeneca. Patients were randomly assigned sequentially in 2 strata (with and without OADs) in balanced block sizes of 4, with the goal of randomly assigning at least 40% of study participants to the insulin-only stratum. A double-dummy technique was used because the 10-mg dapagliflozin tablets were slightly larger than the 2.5- or 5-mg tablets. All placebos were identical in appearance, odor, and taste to their corresponding investigational products.

Patients, investigators, study monitors, and personnel at AstraZeneca and Bristol-Myers Squibb had no access to the randomization scheme, except for cases of medical emergencies. Because primary efficacy analyses were planned at 24 weeks, personnel at AstraZeneca and Bristol-Myers Squibb had access to the data at that time. During the double-blind extension periods, investigators, patients, and study monitors remained blinded, except for cases of medical emergencies.

No dose modifications of study medication or OADs were allowed during the treatment phase, except to decrease doses of OADs when hypoglycemia after cessation of insulin therapy was a concern. Daily insulin doses were held constant (within 10% of the baseline dose) during the study unless changes were needed to ensure patients' well-being. Up-titration of insulin (defined as an increase in daily dose of <5 U and of <10% from baseline) was permitted if fasting blood glucose level was greater than 13.3 mmol/L (>240 mg/dL) between weeks 0 and 12, greater than 12.2 mmol/L (>220 mg/dL) between weeks 12 and 24, or greater than 9.9 mmol/L (>180 mg/dL) between weeks 25 and 48 on at least 3 self-monitored readings in the 7 days before the study visit or on a site-measured reading at the study visit (confirmed by repeated site and central laboratory measurements). Up-titration was also permitted if HbA1c was greater than 8% between weeks 25 and 48. Insulin was down-titrated if 2 or more self-monitored blood glucose readings were 4.4 mmol/L or less (≤80 mg/dL) in the first 7 days of active randomized treatment or 3.8 mmol/L or less (≤70 mg/dL) after the first 7 days.

Patients

Men and women aged 18 to 80 years were enrolled if they had type 2 diabetes mellitus, a body mass index of 45 kg/m2 or less, and inadequate glycemic control (HbA1c ≥7.5% and ≤10.5%). Participants had to have received a stable insulin regimen with a mean daily insulin dose of 30 U or more for at least 8 weeks, with daily insulin requirements varying by more than 10% on no more than 1 occasion in the 7 days before randomization. Additional treatment with stable doses of up to 2 OADs was allowed. Patients had to receive at least 1500 mg daily or their maximum tolerated dose of metformin or at least half of the daily maximum dose of other OADs for at least 8 weeks before enrollment. Patients were instructed to follow a stable diet and exercise regimen after enrollment.

Major exclusion criteria were type 1 diabetes mellitus; symptoms of poorly controlled diabetes; calculated creatinine clearance less than 50 mL/min per 1.73 m2; or a measured serum creatinine level greater than 177 µmol/L (>2 mg/dL) or, if receiving metformin, greater than 133 µmol/L (>1.5 mg/dL) for men and at least 124 µmol/L (≥1.4 mg/dL) for women.

Trial Outcomes

All primary and key secondary efficacy variables were prespecified. The primary outcome was change in HbA1c from baseline to week 24. The 4 key secondary efficacy variables at 24 weeks were change in total body weight from baseline, change in calculated mean daily insulin dose, proportion of patients with calculated mean daily insulin dose reductions of 10% or more from baseline, and change in fasting plasma glucose (determined by 1 central laboratory measurement at each study visit) from baseline. The aim at 48 weeks was to assess the safety and tolerability of dapagliflozin therapy and to explore whether the effects on HbA1c and the secondary efficacy variables were maintained.

Safety evaluations over 48 weeks included adverse events, laboratory variables, and vital signs. Major hypoglycemia was defined as a symptomatic episode in which the patient required external assistance, had a capillary or plasma glucose level less than 3 mmol/L (<54 mg/dL), and promptly recovered after receiving glucose or glucagon. Minor hypoglycemia was defined as a symptomatic episode in which the patient had a capillary or plasma glucose level less than 3.5 mmol/L (<63 mg/dL), regardless of the need for external assistance, or an asymptomatic capillary or plasma glucose level less than 3.5 mmol/L (<63 mg/dL) that did not qualify as a major episode. Other hypoglycemia was defined as a suggestive episode that did not meet the criteria for a major or minor episode. A prespecified list of preferred terms from the Medical Dictionary for Regulatory Activities, version 12.1, identified signs, symptoms, and other reports in the database that suggested genital infection or urinary tract infection (UTI). This list included terms for nonspecific signs and symptoms that suggested genital infection (such as genital or vulvovaginal pruritus) and terms for clinical infection (such as vaginal infection) but did not include terms for sexually transmitted diseases. Patients reported these events both spontaneously and in response to questions proactively asked by the investigator during study visits. Both spontaneous and solicited responses were coded by using the prespecified list of preferred terms.

Statistical Analysis

Two analysis sets were defined: the safety set, comprising all randomly assigned patients who received at least 1 dose of study medication, and the full set, comprising all randomly assigned patients who received at least 1 dose of study medication and had a nonmissing baseline value and at least 1 postbaseline efficacy value for at least 1 efficacy variable.

Efficacy variables were analyzed with the full analysis set. Planned analyses over 48 weeks are reported with data from after insulin up-titration for all variables. Adjusted point estimates and 95% CIs are shown as originally planned. All reported P values are nominal and unadjusted for multiple comparisons.

For continuous variables, a mixed model (using the PROC MIXED procedure in SAS [SAS Institute, Cary, North Carolina]) assessed changes from baseline with fixed effects for treatment group, OAD use, week, baseline value as a covariate, and interactions of week with treatment group and week with the baseline covariate. An unstructured covariance matrix was applied for repeated measures in a patient (19). To probe the validity of the missing-at-random assumption underlying the mixed model and to explore the potential effect of informative missing data, missing not at random, a sensitivity analysis was conducted by using pattern-mixture modeling that assumed control-based pattern imputation (20). Appendix 2 provides further details.

Categorical variables were analyzed by using the method of Zhang and colleagues (21), with adjustment for baseline mean daily dose of insulin and OAD use. Patients with missing data were included in these analyses and considered nonresponders.

The Kaplan–Meier method was used to analyze time to onset of up-titration for not achieving prespecified glycemic targets or discontinuation of therapy due to poor glycemic control.

Analysis of differences in proportions of patients experiencing adverse events of interest between the pooled dapagliflozin groups and the placebo group was performed by using Proc-StatXact 4 (Cytel Software, Cambridge, Massachusetts). All other statistical analyses were done with SAS, version 8.2. The frequency of general adverse events and changes in laboratory variables were summarized by using descriptive statistics.

Sample sizes were calculated on the basis of anticipated differences in the primary efficacy variable. To detect a difference of 0.5% at a 2-sided significance level of 0.019 between dapagliflozin versus placebo for changes in HbA1c level from baseline to week 24 (assuming an SD of 1.2%), 153 patients per group were needed to provide 90% power. Assuming that 5% of patients would not be evaluable, the randomization target was 161 patients per group (a total of 644). An initial enrollment target of 1610 patients was calculated to account for screening failures.

Role of the Funding Source

Our study was sponsored by AstraZeneca and Bristol-Myers Squibb. The sponsors were involved in the study design and the data collection, review, and analysis. The report was prepared by the authors, with editorial assistance funded by the sponsors. All authors had full access to the data, were responsible for interpreting the data, and had final responsibility for the content of the report and the decision to submit the manuscript for publication.

Demographic and Baseline Characteristics

Because screening failure rates were lower than anticipated, 808 of 1240 enrolled patients were randomly assigned before enrollment could be stopped (Figure 1). Demographic and baseline characteristics were similar across all groups (Table 1). The mean duration of insulin therapy was approximately 6 years, about half of the mean time since diagnosis of type 2 diabetes mellitus (13.6 years). Mean daily insulin dose was 77.1 U, with 17% using only long-acting basal insulin and 83% using a sliding scale (bolus) regimen. Fifty percent of patients were receiving other background OADs (principally metformin) in addition to insulin, mean baseline HbA1c level was 8.53%, and mean baseline fasting plasma glucose level was 9.9 mmol/L (177.6 mg/dL).

Grahic Jump Location

Figure 1. Study flow diagram.

DAPA = dapagliflozin; INS = insulin; OAD = oral antidiabetic drug; PLA = placebo.

* This patient received no study medication or postbaseline assessments.

Grahic Jump Location
Table Jump Placeholder Table 1. Major Demographic and Baseline Characteristics
Efficacy Outcomes
Primary Efficacy Variable

The most rapid decreases in mean HbA1c level with dapagliflozin occurred over the first 8 weeks; differences were significantly greater in all dapagliflozin groups than in the placebo group at 24 weeks, and these differences were maintained at 48 weeks (Figure 2, top). Mean HbA1c level in the 10-mg dapagliflozin group was reduced by 0.54% (95% CI, 0.38% to 0.70%; P < 0.001) compared with placebo at 48 weeks.

Grahic Jump Location

Figure 2. Adjusted mean changes in HbA1c level over time ( top) and at 48 weeks (bottom).

Samples are patients in the full analysis set with nonmissing baseline values and nonmissing values for a given time point. Sample sizes at time 0 are 193, 202, 211, and 193 for the PLA, 2.5-mg DAPA, 5-mg DAPA, and 10-mg DAPA groups, respectively. Treatment group symbols are shifted horizontally to prevent the error bars from overlapping. DAPA = dapagliflozin; HbA1c = hemoglobin A1c; INS = insulin; PLA = placebo.

Grahic Jump Location

A sensitivity analysis using a pattern-mixture model (under the assumption that patients receiving dapagliflozin who discontinued the study would exhibit the same future evolution of disease as patients receiving placebo who remained in the study) provided results similar to those from the primary mixed-model analysis of HbA1c (Appendix 2). Alternative means of handling missing data resulted in similar results for the primary efficacy variable (Appendix 3).

An analysis by strata showed an adjusted mean difference in HbA1c at 48 weeks between the 10-mg dapagliflozin and placebo groups of −0.61% (CI, −0.82% to −0.40%) in patients receiving insulin and 1 or 2 OADs, compared with −0.46% (CI, −0.71% to −0.21%) in those receiving insulin alone (Appendix Table 1).

Table Jump Placeholder Appendix Table 1. Analysis by Strata: Adjusted Mean Changes in HbA1c Level at Week 48 in Patients Receiving Insulin Alone or With Other OADs
Key Secondary Efficacy Variables

Steady decreases in body weight occurred over time with dapagliflozin. At 24 weeks, significantly greater differences were found in all dapagliflozin groups than in the placebo group, and these differences were maintained at 48 weeks (Figure 3, top). Mean total body weight decreased by 1.61 kg in the 10-mg dapagliflozin group over 24 weeks, compared with an increase of 0.43 kg in the placebo group (mean difference, −2.04 kg [CI, −2.59 to −1.48 kg]; P < 0.001). At 48 weeks, the mean difference was −2.43 kg (CI, −3.18 to −1.68 kg; P < 0.001).

Grahic Jump Location

Figure 3. Adjusted mean changes in total body weight over time (top) and at 48 weeks (bottom).

Samples are patients in the full analysis set with nonmissing baseline values and nonmissing values for a given time point. Sample sizes at time 0 are 193, 202, 211, and 193 for the PLA, 2.5-mg DAPA, 5-mg DAPA, and 10-mg DAPA groups, respectively. Treatment group symbols are shifted horizontally to prevent the error bars from overlapping. DAPA = dapagliflozin; INS = insulin; PLA = placebo.

Grahic Jump Location

Dapagliflozin therapy did not increase mean daily insulin requirements over 48 weeks, but these requirements increased progressively in the placebo group (Appendix Figure 1, top), resulting in a total increase of 10.54 U at 48 weeks (Appendix Figure 1, bottom). At 24 weeks, the adjusted proportion of patients with a mean daily insulin dose reduction of at least 10% from baseline was 10.2% in the placebo group; this proportion differed from the placebo group by 6.3 percentage points (CI, −0.4 to 13.0 percentage points; P = 0.064) in the 2.5-mg dapagliflozin group, 6.3 percentage points (CI, −0.3 to 12.9 percentage points; P = 0.060) in the 5-mg dapagliflozin group, and 8.9 percentage points (CI, 1.9 to 15.9 percentage points; P = 0.013) in the 10-mg dapagliflozin group. At 48 weeks, the adjusted proportion of patients with a mean daily insulin dose reduction of at least 10% from baseline was 10.5% in the placebo group; this proportion differed from the placebo group by 7.6 percentage points (CI, 0.8 to 14.5 percentage points; P = 0.030) in the 2.5-mg dapagliflozin group, 7.0 percentage points (CI, 0.3 to 13.7 percentage points; P = 0.041) in the 5-mg dapagliflozin group, and 8.1 percentage points (CI, 1.1 to 15.1 percentage points; P = 0.024) in the 10-mg dapagliflozin group.

Grahic Jump Location

Appendix Figure 1. Adjusted mean changes in daily insulin dose over time (top) and at 48 weeks (bottom).

Samples are patients in the full analysis set with nonmissing baseline values and nonmissing values for a given time point. Sample sizes at time 0 are 191, 200, 209, and 194 for the PLA, 2.5-mg DAPA, 5-mg DAPA, and 10-mg DAPA groups, respectively. Treatment group symbols are shifted horizontally to prevent the error bars from overlapping. DAPA = dapagliflozin; INS = insulin; PLA = placebo.

Grahic Jump Location

The differences in adjusted mean change in fasting plasma glucose level from baseline at 24 weeks were −0.65 mmol/L (CI, −1.19 to −0.11 mmol/L) in the 2.5-mg dapagliflozin group, −1.12 mmol/L (CI, −1.66 to −0.59 mmol/L) in the 5-mg dapagliflozin group, and −1.10 mmol/L (CI, −1.64 to −0.56 mmol/L) in the 10-mg dapagliflozin group (all P < 0.001). At 48 weeks, these values were −0.69 mmol/L (CI, −1.28 to −0.11 mmol/L), −0.90 mmol/L (CI, −1.48 to −0.33 mmol/L), and −0.94 mmol/L (CI, −1.53 to −0.36 mmol/L), respectively (all P < 0.001). (To convert fasting plasma glucose values from mmol/L to mg/dL, divide by 0.0555.)

After 24 weeks, 9.7% to 11.2% of patients receiving dapagliflozin required insulin up-titration (an increase in daily dose of >5 U and of >10% from baseline) or discontinued the study because of poor glycemic control, compared with 29.2% of patients receiving placebo. Most of these differences were accounted for by insulin up-titration, with only 2 discontinuations due to poor control by week 24 (1 in the 2.5-mg dapagliflozin group and the other in the placebo group). Over 48 weeks, the differences between the placebo and dapagliflozin groups increased further (Appendix Figure 2 and Appendix Table 2).

Grahic Jump Location

Appendix Figure 2. Probability of insulin up-titration due to not achieving prespecified glycemic targets or of discontinuation due to poor glycemic control over time.

Symbols represent censored observations. “Week” is the actual number of days from the first dose of double-blind study medication divided by 7, not the scheduled visit week. Values for patients at risk are for the beginning of the period. DAPA = dapagliflozin; INS = insulin; PLA = placebo.

Grahic Jump Location
Table Jump Placeholder Appendix Table 2. Patients Receiving Insulin Up-Titration or Discontinuing Therapy for Not Achieving Prespecified Glycemic Targets
Vital Signs, Safety, and Adverse Events Over 48 Weeks

Greater mean decreases from baseline in seated systolic blood pressure were observed in some dapagliflozin groups compared with the placebo group. Mean change in seated diastolic blood pressure and heart rate from baseline to 24 and 48 weeks did not systematically differ between the dapagliflozin and placebo groups. No increase in the proportion of patients who had orthostatic hypotension was noted in the dapagliflozin groups (Appendix Table 3).

Table Jump Placeholder Appendix Table 3. Adjusted Mean Changes in Vital Signs at Weeks 24 and 48

Similar proportions of patients experienced adverse events over 48 weeks across all groups (Table 2). Higher proportions of patients in the dapagliflozin groups had adverse events that were considered treatment-related by a blinded investigator (21.3% in the 2.5-mg dapagliflozin group, 29.2% in the 5-mg dapagliflozin group, and 29.1% in the 10-mg dapagliflozin group) than in the placebo group (20.8%). Proportions of patients with serious adverse events were similar across groups (13.2% in the placebo group and ≤13.4% in the dapagliflozin groups). Three serious adverse events were considered treatment-related by the investigator (1 patient with hypoglycemia and 1 with a change of bowel habit, both in the 5-mg dapagliflozin group, and 1 patient with constipation in the 10-mg dapagliflozin group). Two deaths occurred in the 5-mg dapagliflozin group (1 patient died of cardiogenic shock 2 days after receiving aortic valve replacement and coronary artery bypass graft surgery, 13 days after dapagliflozin therapy had been stopped, and 1 patient died of acute myocardial infarction).

Table Jump Placeholder Table 2. Patients With Adverse Events

Higher proportions of patients had at least 1 hypoglycemic event in the dapagliflozin groups than in the placebo group (60.4%, 55.7%, and 53.6% in the 2.5-mg, 5-mg, and 10-mg dapagliflozin groups, respectively, vs. 51.8%) (Table 2). Three patients experienced serious adverse events due to hypoglycemia: 2 patients in the 5-mg dapagliflozin group had events that the study investigator considered to be serious, and 1 patient in the placebo group had hypoglycemic coma. No patient discontinued the study because of hypoglycemia.

Events suggesting genital infection or UTI were higher in all dapagliflozin groups than in the placebo group (Table 2). The difference in proportions of patients between the pooled dapagliflozin groups and the placebo group was significant for genital infection (6.5 percentage points [CI, 1.2 to 12.4 percentage points]) but not for UTI (4.6 percentage points [CI, −1.3 to 11.1 percentage points]). These events were more common in women and mostly occurred during the first 24 weeks of treatment (Appendix Figure 3). Most suggestive events were classified as mild or moderate in intensity and responded to routine management. Two patients in the 10-mg dapagliflozin group discontinued treatment because of genital infection before week 24; 2 patients in the 5-mg dapagliflozin group discontinued treatment because of lower UTI, 1 before week 24 and 1 during weeks 24 to 48; and 1 patient in the 2.5-mg dapagliflozin group discontinued treatment because of pyelonephritis before week 24.

Grahic Jump Location

Appendix Figure 3. Patients with events suggesting genital infection (top) or urinary tract infection (bottom) at 24 and 48 weeks.

DAPA = dapagliflozin; INS = insulin; PLA = placebo.

Grahic Jump Location

Clinically meaningful increases were observed in urinary glucose, hematocrit, serum creatinine, blood urea nitrogen, and cystatin C levels and decreases in serum uric acid levels and calculated creatinine clearance, with greater absolute changes in the dapagliflozin groups (Appendix Table 4). However, these changes were not accompanied by increased rates of renal impairment or failure, hypotension, dehydration, or hypovolemia (Table 2).

Table Jump Placeholder Appendix Table 4. Change in Laboratory Values of Interest From Baseline to Week 48

Many patients with type 2 diabetes treated with OADs eventually require insulin to manage progressive deterioration in glycemic control (3). However, weight gain and peripheral edema commonly occur with escalating insulin therapy, reducing patient satisfaction and treatment adherence (22). In our 48-week study, dapagliflozin improved glycemic control, prevented insulin dose escalation, and produced significant and sustained weight loss.

To better comprehend the effects of adding dapagliflozin therapy to preexisting background insulin therapy, we did not reduce insulin doses before starting trial therapy, in contrast to the previous dapagliflozin pilot study (18) that reported lower rates of hypoglycemia and greater weight loss. In our study, more minor hypoglycemic episodes were noted in the dapagliflozin groups than in the placebo group, and a higher proportion of patients who received dapagliflozin required insulin down-titration of more than 10%. An increase in hypoglycemia is commonly observed when agents with a low intrinsic hypoglycemic propensity are added to insulin (23) or sulfonylurea (24) therapy.

Diabetes is associated with increased genital infection and UTI in women (2526), possibly due to poor glycemic control, which includes hyperglycemia as well as glucosuria (27). Because the mechanism of action of dapagliflozin involves increased urinary glucose excretion, monitoring rates of genital infection and UTI in clinical trials is important. Treatment with dapagliflozin was associated with increased signs, symptoms, and reports suggesting genital infection and UTI, predominantly during the first 24 weeks of the study, although many reported instances of UTI were not confirmed with urine cultures. This aspect of dapagliflozin use requires further investigation, but it is reassuring to note that these events responded to conventional management and rarely led to study withdrawals.

Diabetes also increases risk for cancer of the liver, pancreas, endometrium, colon, rectum, breast, and bladder (2829). Dapagliflozin has not demonstrated genotoxicity in animal studies, and no increase in tumors has been observed over 2 years in carcinogenicity studies in rodents (30) when mice and rats were exposed to doses 105 and 186 times higher, respectively, than the 10-mg daily dose used in humans. Consistent with these preclinical safety data, no overall imbalance in cases of cancer has been observed between the 4459 patients who received dapagliflozin and the 2239 control patients from the dapagliflozin clinical development program. Although some types of cancer were overrepresented in control patients, numerical imbalances were noted in occurrences of breast and bladder cancer in patients who received dapagliflozin (30), including 3 cases of each of these types of cancer in the dapagliflozin groups from our study after 2 years of follow-up. Dapagliflozin is highly selective for sodium–glucose cotransporter 2, which is not known to be expressed in either bladder or breast tissue. The mechanism of action of dapagliflozin has no known link to tumor risk; however, long-term surveillance will be required to exclude any potential association.

Our study has limitations. First, insulin doses were not titrated to target. However, our study was designed to demonstrate whether adding dapagliflozin to insulin therapy can improve glycemic control, requiring baseline therapy to remain unchanged, unless deteriorating glycemic control dictates the need for insulin up- or down-titration according to predefined study criteria. Second, because more appropriate analytic methods than those that were initially planned were used, our P values are considered post hoc, nominal, and unadjusted for multiple comparisons. Third, approximately 95% of recruited patients were white, although racial origin is unlikely to affect the mechanism of action of dapagliflozin. Finally, we could not fully evaluate potential concerns about risk for breast or bladder cancer with dapagliflozin because of the lack of statistical power in a study of this sample size and duration.

In conclusion, in patients whose type 2 diabetes mellitus was inadequately controlled with insulin therapy, adding dapagliflozin reduced HbA1c levels and weight over 48 weeks without increasing overall insulin dose; however, increased rates of minor hypoglycemia were noted. Conversely, patients for whom placebo was added to insulin therapy had a progressive increase in insulin dose and weight. These data suggest that dapagliflozin may offer a new treatment option for patients receiving insulin therapy whose type 2 diabetes remains inadequately controlled.

Hoerger TJ, Segel JE, Gregg EW, Saaddine JB.  Is glycemic control improving in U.S. adults? Diabetes Care. 2008; 31.81-6
PubMed
CrossRef
 
Hsu WC.  Consequences of delaying progression to optimal therapy in patients with type 2 diabetes not achieving glycemic goals. South Med J. 2009; 102.67-76
PubMed
CrossRef
 
Buse J.  Combining insulin and oral agents. Am J Med. 2000; 108.6a23S-32S
PubMed
CrossRef
 
Holman RR, Thorne KI, Farmer AJ, Davies MJ, Keenan JF, Paul S, et al. 4-T Study Group.  Addition of biphasic, prandial, or basal insulin to oral therapy in type 2 diabetes. N Engl J Med. 2007; 357.1716-30
PubMed
CrossRef
 
Nathan DM, Buse JB, Davidson MB, Ferrannini E, Holman RR, Sherwin R, et al. American Diabetes Association.  Medical management of hyperglycemia in type 2 diabetes: a consensus algorithm for the initiation and adjustment of therapy: a consensus statement of the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes Care. 2009; 32.193-203
PubMed
CrossRef
 
UK Prospective Diabetes Study (UKPDS) Group.  Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet. 1998; 352.837-53
PubMed
CrossRef
 
Yki-Järvinen H.  Combination therapies with insulin in type 2 diabetes. Diabetes Care. 2001; 24.758-67
PubMed
CrossRef
 
Holman RR, Farmer AJ, Davies MJ, Levy JC, Darbyshire JL, Keenan JF, et al. 4-T Study Group.  Three-year efficacy of complex insulin regimens in type 2 diabetes. N Engl J Med. 2009; 361.1736-47
PubMed
CrossRef
 
Wright EM, Hirayama BA, Loo DF.  Active sugar transport in health and disease. J Intern Med. 2007; 261.32-43
PubMed
CrossRef
 
Han S, Hagan DL, Taylor JR, Xin L, Meng W, Biller SA. et al.  Dapagliflozin, a selective SGLT2 inhibitor, improves glucose homeostasis in normal and diabetic rats. Diabetes. 2008; 57.1723-9
PubMed
CrossRef
 
Meng W, Ellsworth BA, Nirschl AA, McCann PJ, Patel M, Girotra RN. 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; 51.1145-9
PubMed
CrossRef
 
Komoroski B, Vachharajani N, Boulton D, Kornhauser D, Geraldes M, Li L. et al.  Dapagliflozin, a novel SGLT2 inhibitor, induces dose-dependent glucosuria in healthy subjects. Clin Pharmacol Ther. 2009; 85.520-6
PubMed
CrossRef
 
Ferrannini E, Ramos SJ, Salsali A, Tang W, List JF.  Dapagliflozin monotherapy in type 2 diabetic patients with inadequate glycemic control by diet and exercise: a randomized, double-blind, placebo-controlled, phase 3 trial. Diabetes Care. 2010; 33.2217-24
PubMed
CrossRef
 
Komoroski B, Vachharajani N, Feng Y, Li L, Kornhauser D, Pfister M.  Dapagliflozin, a novel, selective SGLT2 inhibitor, improved glycemic control over 2 weeks in patients with type 2 diabetes mellitus. Clin Pharmacol Ther. 2009; 85.513-9
PubMed
CrossRef
 
List JF, Woo V, Morales E, Tang W, Fiedorek FT.  Sodium-glucose cotransport inhibition with dapagliflozin in type 2 diabetes. Diabetes Care. 2009; 32.650-7
PubMed
CrossRef
 
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
PubMed
CrossRef
 
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
PubMed
CrossRef
 
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
PubMed
CrossRef
 
Verbeke G, Molenberghs G.  Linear Mixed Models for Longitudinal Data. New York: Springer-Verlag; 2000.
 
Ratitch B, O'Kelly M.  Implementation of Pattern-Mixture Models Using Standard SAS/STAT Procedures. Saint-Laurent, Quebec, Canada: Quintiles; 2011. Accessed at www.quintiles.com/elements/media/inthenews/implementation-pattern-mixture-models-using-standard-sasstat-procedures.pdf on 24 January 2012.
 
Zhang M, Tsiatis AA, Davidian M.  Improving efficiency of inferences in randomized clinical trials using auxiliary covariates. Biometrics. 2008; 64.707-15
PubMed
CrossRef
 
Pi-Sunyer FX.  The impact of weight gain on motivation, compliance, and metabolic control in patients with type 2 diabetes mellitus. Postgrad Med. 2009; 121.94-107
PubMed
CrossRef
 
Charbonnel B, DeFronzo R, Davidson J, Schmitz O, Birkeland K, Pirags V, et al. PROactive investigators.  Pioglitazone use in combination with insulin in the prospective pioglitazone clinical trial in macrovascular events study (PROactive19). J Clin Endocrinol Metab. 2010; 95.2163-71
PubMed
CrossRef
 
Hermansen K, Kipnes M, Luo E, Fanurik D, Khatami H, Stein P, Sitagliptin Study 035 Group.  Efficacy and safety of the dipeptidyl peptidase-4 inhibitor, sitagliptin, in patients with type 2 diabetes mellitus inadequately controlled on glimepiride alone or on glimepiride and metformin. Diabetes Obes Metab. 2007; 9.733-45
PubMed
CrossRef
 
Foxman B.  Epidemiology of urinary tract infections: incidence, morbidity, and economic costs. Dis Mon. 2003; 49.53-70
PubMed
CrossRef
 
Shah BR, Hux JE.  Quantifying the risk of infectious diseases for people with diabetes. Diabetes Care. 2003; 26.510-3
PubMed
CrossRef
 
Bohannon NJ.  Treatment of vulvovaginal candidiasis in patients with diabetes. Diabetes Care. 1998; 21.451-6
PubMed
CrossRef
 
Vigneri P, Frasca F, Sciacca L, Pandini G, Vigneri R.  Diabetes and cancer. Endocr Relat Cancer. 2009; 16.1103-23
PubMed
CrossRef
 
Larsson SC, Orsini N, Brismar K, Wolk A.  Diabetes mellitus and risk of bladder cancer: a meta-analysis. Diabetologia. 2006; 49.2819-23
PubMed
CrossRef
 
Bristol-Myers Squibb, AstraZeneca.  U.S. Food and Drug Administration Endocrinologic & Metabolic Advisory Committee Background Document: Dapagliflozin, BMS-512148, NDA 202293. Princeton, NJ: Bristol-Myers Squibb: 2011. Accessed at www.fda.gov/AdvisoryCommittees/CommitteesMeetingMaterials/Drugs/EndocrinologicandMetabolicDrugsAdvisoryCommittee/ucm262993.htm on 24 January 2012.
 
National Research Council.  The Prevention and Treatment of Missing Data in Clinical Trials. Washington, DC: National Academies Pr; 2010.
 
Beunckens C, Molenberghs G, Kenward MG.  Direct likelihood analysis versus simple forms of imputation for missing data in randomized clinical trials. Clin Trials. 2005; 2.379-86
PubMed
CrossRef
 
Hamer RM, Simpson PM.  Last observation carried forward versus mixed models in the analysis of psychiatric clinical trials [Editorial]. Am J Psychiatry. 2009; 166.639-41
PubMed
CrossRef
 
Appendix 1: Dapagliflozin 006 Study Group Investigators

Austria: Hoppichler F, Luger A, Prager R, Roden M, Kotter T, and Schernthaner G.

Bulgaria: Borisova A, Daskalova I, Hristozov K, Lazarova G, Orbetzova M, Petkova M, Slavcheva A, Stoykova Y, Tankova C, Veleva N, and Zaharieva S.

Canada: Ardilouze J, Aronson R, Boucher P, Carthy J, Dzongowski P, Gaudet D, Godsell S, Goldenberg R, Hart R, Houlden R, Kaiser S, Khandwala H, Landry D, Lau D, O'Keefe D, Sigalas J, Tellier G, Twum-Barima Y, and Woo V.

Finland: Alanko J, Kurl S, Mäkelä J, Niskanen L, Paul R, Perhonen M, Pihlman S, Rissanen A, Strand J, Valle T; Germany: Bieler T, Fuchs F, Götze D, Hanefeld M, Mölle A, Paschen B, Rose L, Schumacher M, Semmler S, Sehnert W, Tillenburg B, Wendisch U, and Winkelmann B.

Great Britain: Bain S, Farmer I, Jassel GS, Owen D, Reed R, Simpson H, Strang C, and Wilding J.

Hungary: Deák L, Gurzó M, Hegyi I, Kerényi Z, Lászlóczky Á, Páll K, Pénzes J, Reiber I, Révész K, Somogyi A, Tamás G, and Vándorfi G.

The Netherlands: Castro Cabezas M, de Koning E, Timmerman R, and van de Wiel A.

Romania: Crisan C, Ferariu I, and Vlaiculescu M.

Russian Federation: Arutyunov G, Dreval A, Grineva E, Karpova I, Kobalava Z, Krasilnikova E, Kukharchuk V, Lantseva O, Pavlova M, Privalov D, Strongin L, Tereschenko S, Vorokhobina N, and Zhelninova T.

Slovakia: Belesova K, Fabry J, Kupcova T, Macko M, Paulovic V, Tomasova L, Truban J, and Vargova A.

Spain: Calle Pascual A, De Teresa Parreño L, Durán García S, Freixenet N, Gomis de Barbará R, González Albarrán O, and González Clemente JM.

United States: Bernstein R, Brusco O, Graf R, Freeman J, Hoekstra J, Khairi R, Lubin B, Norwood P, Reed J, Rosenstock J, Soler N, Sugimoto D, and Wahl T.

Appendix 2: Sensitivity Analysis of the Primary Efficacy Variable Using Pattern-Mixture Modeling

To probe the validity of the missing-at-random (MAR) assumption underlying the mixed model and to explore the potential effect of informative missing data, missing not at random (MNAR), we used pattern-mixture modeling to conduct a sensitivity analysis that assumed control-based pattern imputation—the assumption that after study discontinuation, patients who discontinue therapy in the experimental treatment groups will exhibit the same future evolution of the disease as patients receiving placebo therapy who continue in the study. Patients who discontinue therapy in the placebo group are assumed to evolve in the same way as patients receiving placebo therapy who continue in the study. The analysis was conducted with the safety analysis set by using HbA1c. Most missing data in longitudinal clinical trials result from patients discontinuing therapy at a particular time point and subsequent data being unavailable; this is known as monotone missing data. A small amount of missing data may result from patients who skip a visit but return for subsequent visits, known as nonmonotone missing data. The pattern-mixture modeling method assumes monotone missing data. Thus, Markov-chain Monte Carlo multiple imputation was used as a preliminary step to generate 100 data sets that contained imputed values for nonmonotone missing data. Each of the resulting data sets, now containing exclusively monotone missing data, was subjected to further stepwise imputation resulting in 100 fully imputed data sets. At each step, monotone missing values at time t were imputed by fitting a regression model that estimated the relationship between the values at week t and week t − 1 on the basis of the available cases in a selected treatment group. Mean imputed HbA1c was compared graphically among patients who completed versus those who dropped out of the study (Appendix Figure 4), and an analysis of covariance (ANCOVA) model of imputed mean change in HbA1c from baseline was conducted by using the MIANALYZE procedure in SAS and compared with the results obtained from the primary mixed-model analysis (Appendix Table 5).

Grahic Jump Location

Appendix Figure 4. Change in mean HbA1c level over time calculated by using the placebo group to impute missing data in the safety analysis set and analyzed by treatment group and discontinuation status.

DAPA = dapagliflozin; HbA1c = hemoglobin A1c; INS = insulin; PLA = placebo.

Grahic Jump Location
Table Jump Placeholder Appendix Table 5. Patients in the Safety Analysis Set With Total, Monotone, and Nonmonotone Missing Data

The difference in the adjusted mean change in imputed HbA1c between the 10-mg dapagliflozin group and the placebo group according to the ANCOVA model (−0.50% [CI, −0.68% to −0.32%]) was similar to that obtained from the mixed-model analysis (−0.54% [CI, −0.70% to −0.38%]), indicating that missing data are unlikely to have substantially affected our assessment of dapagliflozin efficacy for the primary variable (Appendix Table 6).

Table Jump Placeholder Appendix Table 6. Adjusted Mean Differences in Change in Hemoglobin A1c Level at Week 48
Appendix 3: Alternative Means of Handling Missing Data

The primary efficacy variable was tested by using the Dunnett method with an α of 0.019 for each pairwise group comparison of dapagliflozin versus placebo (overall α level = 0.05) by using ANCOVA with treatment group and OAD use as fixed effects and baseline value as a covariate in the full analysis set. Data after insulin up-titration were excluded, and these and other missing values were replaced by using the last-observation-carried-forward (LOCF) method (Appendix Table 7).

Table Jump Placeholder Appendix Table 7. Protocol-Prespecified Analysis of the Primary Efficacy Variable at 24 Weeks

Although our analysis was prespecified, it is important to note that recent U.S. Committee on National Statistics recommendations (31) for the prevention and treatment of missing data in clinical trials have stated that LOCF analyses are problematic for the following reasons (32).

First, a necessary but insufficient assumption required by the LOCF method is that data are missing completely at random (for example, a specimen is accidentally dropped in the laboratory). In practice, this assumption is rarely valid and data are more commonly either MAR (for example, a patient withdraws from a study because he or she does not achieve a predefined threshold for treatment efficacy) or MNAR (for example, a patient drops out of the study because of an unforeseen adverse change in his or her health status). Second, the LOCF method further assumes that patients' responses remain unchanged from the last observed value to the end point of the trial and is regarded by some as a conservative estimate of treatment effect. However, these conditions are seldom valid because patients could have either improved or deteriorated had they remained in the trial. This may lead to under- or overestimation of treatment effects when using the LOCF method (32). Alternative models that are valid under the MAR assumption, such as repeated measures mixed-effects analyses, can reduce bias due to dropout by using data before dropout to project individual-patient regression lines to the end of the study, rather than by simply projecting the last observed value to the end of the study (as with the LOCF method) (33). Finally, carrying values forward from an early point in the trial can lead to underestimation of the associated SEs, artificially increasing the amount of information in the data by treating imputed and actually observed values on equal footing, which may result in an erroneous declaration of statistical significance (32).

In addition, analyses that exclude data on the basis of postrandomization events are subject to bias. We therefore used more appropriate means of analysis in our report: Mixed-effects models valid under MAR data with an exploration of the potential effect of MNAR data probed by using pattern-mixture model sensitivity analyses (31).

Figures

Grahic Jump Location

Figure 1. Study flow diagram.

DAPA = dapagliflozin; INS = insulin; OAD = oral antidiabetic drug; PLA = placebo.

* This patient received no study medication or postbaseline assessments.

Grahic Jump Location
Grahic Jump Location

Figure 2. Adjusted mean changes in HbA1c level over time ( top) and at 48 weeks (bottom).

Samples are patients in the full analysis set with nonmissing baseline values and nonmissing values for a given time point. Sample sizes at time 0 are 193, 202, 211, and 193 for the PLA, 2.5-mg DAPA, 5-mg DAPA, and 10-mg DAPA groups, respectively. Treatment group symbols are shifted horizontally to prevent the error bars from overlapping. DAPA = dapagliflozin; HbA1c = hemoglobin A1c; INS = insulin; PLA = placebo.

Grahic Jump Location
Grahic Jump Location

Figure 3. Adjusted mean changes in total body weight over time (top) and at 48 weeks (bottom).

Samples are patients in the full analysis set with nonmissing baseline values and nonmissing values for a given time point. Sample sizes at time 0 are 193, 202, 211, and 193 for the PLA, 2.5-mg DAPA, 5-mg DAPA, and 10-mg DAPA groups, respectively. Treatment group symbols are shifted horizontally to prevent the error bars from overlapping. DAPA = dapagliflozin; INS = insulin; PLA = placebo.

Grahic Jump Location
Grahic Jump Location

Appendix Figure 1. Adjusted mean changes in daily insulin dose over time (top) and at 48 weeks (bottom).

Samples are patients in the full analysis set with nonmissing baseline values and nonmissing values for a given time point. Sample sizes at time 0 are 191, 200, 209, and 194 for the PLA, 2.5-mg DAPA, 5-mg DAPA, and 10-mg DAPA groups, respectively. Treatment group symbols are shifted horizontally to prevent the error bars from overlapping. DAPA = dapagliflozin; INS = insulin; PLA = placebo.

Grahic Jump Location
Grahic Jump Location

Appendix Figure 2. Probability of insulin up-titration due to not achieving prespecified glycemic targets or of discontinuation due to poor glycemic control over time.

Symbols represent censored observations. “Week” is the actual number of days from the first dose of double-blind study medication divided by 7, not the scheduled visit week. Values for patients at risk are for the beginning of the period. DAPA = dapagliflozin; INS = insulin; PLA = placebo.

Grahic Jump Location
Grahic Jump Location

Appendix Figure 3. Patients with events suggesting genital infection (top) or urinary tract infection (bottom) at 24 and 48 weeks.

DAPA = dapagliflozin; INS = insulin; PLA = placebo.

Grahic Jump Location
Grahic Jump Location

Appendix Figure 4. Change in mean HbA1c level over time calculated by using the placebo group to impute missing data in the safety analysis set and analyzed by treatment group and discontinuation status.

DAPA = dapagliflozin; HbA1c = hemoglobin A1c; INS = insulin; PLA = placebo.

Grahic Jump Location

Tables

Table Jump Placeholder Table 1. Major Demographic and Baseline Characteristics
Table Jump Placeholder Appendix Table 1. Analysis by Strata: Adjusted Mean Changes in HbA1c Level at Week 48 in Patients Receiving Insulin Alone or With Other OADs
Table Jump Placeholder Appendix Table 2. Patients Receiving Insulin Up-Titration or Discontinuing Therapy for Not Achieving Prespecified Glycemic Targets
Table Jump Placeholder Appendix Table 3. Adjusted Mean Changes in Vital Signs at Weeks 24 and 48
Table Jump Placeholder Table 2. Patients With Adverse Events
Table Jump Placeholder Appendix Table 4. Change in Laboratory Values of Interest From Baseline to Week 48
Table Jump Placeholder Appendix Table 5. Patients in the Safety Analysis Set With Total, Monotone, and Nonmonotone Missing Data
Table Jump Placeholder Appendix Table 6. Adjusted Mean Differences in Change in Hemoglobin A1c Level at Week 48
Table Jump Placeholder Appendix Table 7. Protocol-Prespecified Analysis of the Primary Efficacy Variable at 24 Weeks

References

Hoerger TJ, Segel JE, Gregg EW, Saaddine JB.  Is glycemic control improving in U.S. adults? Diabetes Care. 2008; 31.81-6
PubMed
CrossRef
 
Hsu WC.  Consequences of delaying progression to optimal therapy in patients with type 2 diabetes not achieving glycemic goals. South Med J. 2009; 102.67-76
PubMed
CrossRef
 
Buse J.  Combining insulin and oral agents. Am J Med. 2000; 108.6a23S-32S
PubMed
CrossRef
 
Holman RR, Thorne KI, Farmer AJ, Davies MJ, Keenan JF, Paul S, et al. 4-T Study Group.  Addition of biphasic, prandial, or basal insulin to oral therapy in type 2 diabetes. N Engl J Med. 2007; 357.1716-30
PubMed
CrossRef
 
Nathan DM, Buse JB, Davidson MB, Ferrannini E, Holman RR, Sherwin R, et al. American Diabetes Association.  Medical management of hyperglycemia in type 2 diabetes: a consensus algorithm for the initiation and adjustment of therapy: a consensus statement of the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes Care. 2009; 32.193-203
PubMed
CrossRef
 
UK Prospective Diabetes Study (UKPDS) Group.  Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet. 1998; 352.837-53
PubMed
CrossRef
 
Yki-Järvinen H.  Combination therapies with insulin in type 2 diabetes. Diabetes Care. 2001; 24.758-67
PubMed
CrossRef
 
Holman RR, Farmer AJ, Davies MJ, Levy JC, Darbyshire JL, Keenan JF, et al. 4-T Study Group.  Three-year efficacy of complex insulin regimens in type 2 diabetes. N Engl J Med. 2009; 361.1736-47
PubMed
CrossRef
 
Wright EM, Hirayama BA, Loo DF.  Active sugar transport in health and disease. J Intern Med. 2007; 261.32-43
PubMed
CrossRef
 
Han S, Hagan DL, Taylor JR, Xin L, Meng W, Biller SA. et al.  Dapagliflozin, a selective SGLT2 inhibitor, improves glucose homeostasis in normal and diabetic rats. Diabetes. 2008; 57.1723-9
PubMed
CrossRef
 
Meng W, Ellsworth BA, Nirschl AA, McCann PJ, Patel M, Girotra RN. 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; 51.1145-9
PubMed
CrossRef
 
Komoroski B, Vachharajani N, Boulton D, Kornhauser D, Geraldes M, Li L. et al.  Dapagliflozin, a novel SGLT2 inhibitor, induces dose-dependent glucosuria in healthy subjects. Clin Pharmacol Ther. 2009; 85.520-6
PubMed
CrossRef
 
Ferrannini E, Ramos SJ, Salsali A, Tang W, List JF.  Dapagliflozin monotherapy in type 2 diabetic patients with inadequate glycemic control by diet and exercise: a randomized, double-blind, placebo-controlled, phase 3 trial. Diabetes Care. 2010; 33.2217-24
PubMed
CrossRef
 
Komoroski B, Vachharajani N, Feng Y, Li L, Kornhauser D, Pfister M.  Dapagliflozin, a novel, selective SGLT2 inhibitor, improved glycemic control over 2 weeks in patients with type 2 diabetes mellitus. Clin Pharmacol Ther. 2009; 85.513-9
PubMed
CrossRef
 
List JF, Woo V, Morales E, Tang W, Fiedorek FT.  Sodium-glucose cotransport inhibition with dapagliflozin in type 2 diabetes. Diabetes Care. 2009; 32.650-7
PubMed
CrossRef
 
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
PubMed
CrossRef
 
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
PubMed
CrossRef
 
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
PubMed
CrossRef
 
Verbeke G, Molenberghs G.  Linear Mixed Models for Longitudinal Data. New York: Springer-Verlag; 2000.
 
Ratitch B, O'Kelly M.  Implementation of Pattern-Mixture Models Using Standard SAS/STAT Procedures. Saint-Laurent, Quebec, Canada: Quintiles; 2011. Accessed at www.quintiles.com/elements/media/inthenews/implementation-pattern-mixture-models-using-standard-sasstat-procedures.pdf on 24 January 2012.
 
Zhang M, Tsiatis AA, Davidian M.  Improving efficiency of inferences in randomized clinical trials using auxiliary covariates. Biometrics. 2008; 64.707-15
PubMed
CrossRef
 
Pi-Sunyer FX.  The impact of weight gain on motivation, compliance, and metabolic control in patients with type 2 diabetes mellitus. Postgrad Med. 2009; 121.94-107
PubMed
CrossRef
 
Charbonnel B, DeFronzo R, Davidson J, Schmitz O, Birkeland K, Pirags V, et al. PROactive investigators.  Pioglitazone use in combination with insulin in the prospective pioglitazone clinical trial in macrovascular events study (PROactive19). J Clin Endocrinol Metab. 2010; 95.2163-71
PubMed
CrossRef
 
Hermansen K, Kipnes M, Luo E, Fanurik D, Khatami H, Stein P, Sitagliptin Study 035 Group.  Efficacy and safety of the dipeptidyl peptidase-4 inhibitor, sitagliptin, in patients with type 2 diabetes mellitus inadequately controlled on glimepiride alone or on glimepiride and metformin. Diabetes Obes Metab. 2007; 9.733-45
PubMed
CrossRef
 
Foxman B.  Epidemiology of urinary tract infections: incidence, morbidity, and economic costs. Dis Mon. 2003; 49.53-70
PubMed
CrossRef
 
Shah BR, Hux JE.  Quantifying the risk of infectious diseases for people with diabetes. Diabetes Care. 2003; 26.510-3
PubMed
CrossRef
 
Bohannon NJ.  Treatment of vulvovaginal candidiasis in patients with diabetes. Diabetes Care. 1998; 21.451-6
PubMed
CrossRef
 
Vigneri P, Frasca F, Sciacca L, Pandini G, Vigneri R.  Diabetes and cancer. Endocr Relat Cancer. 2009; 16.1103-23
PubMed
CrossRef
 
Larsson SC, Orsini N, Brismar K, Wolk A.  Diabetes mellitus and risk of bladder cancer: a meta-analysis. Diabetologia. 2006; 49.2819-23
PubMed
CrossRef
 
Bristol-Myers Squibb, AstraZeneca.  U.S. Food and Drug Administration Endocrinologic & Metabolic Advisory Committee Background Document: Dapagliflozin, BMS-512148, NDA 202293. Princeton, NJ: Bristol-Myers Squibb: 2011. Accessed at www.fda.gov/AdvisoryCommittees/CommitteesMeetingMaterials/Drugs/EndocrinologicandMetabolicDrugsAdvisoryCommittee/ucm262993.htm on 24 January 2012.
 
National Research Council.  The Prevention and Treatment of Missing Data in Clinical Trials. Washington, DC: National Academies Pr; 2010.
 
Beunckens C, Molenberghs G, Kenward MG.  Direct likelihood analysis versus simple forms of imputation for missing data in randomized clinical trials. Clin Trials. 2005; 2.379-86
PubMed
CrossRef
 
Hamer RM, Simpson PM.  Last observation carried forward versus mixed models in the analysis of psychiatric clinical trials [Editorial]. Am J Psychiatry. 2009; 166.639-41
PubMed
CrossRef
 

Letters

NOTE:
Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).

Comments

Submit a Comment
Submit a Comment

Summary for Patients

Dapagliflozin Treatment in Patients With Inadequately Controlled Type 2 Diabetes Despite the Use of Insulin

Summaries for Patients are a service provided by Annals to help patients better understand the complicated and often mystifying language of modern medicine.

Summaries for Patients are presented for informational purposes only. These summaries are not a substitute for advice from your own medical provider. If you have questions about this material, or need medical advice about your own health or situation, please contact your physician. The summaries may be reproduced for not-for-profit educational purposes only. Any other uses must be approved by the American College of Physicians.

Read More...

Clinical Slide Sets

Terms of Use

The In the Clinic® slide sets are owned and copyrighted by the American College of Physicians (ACP). All text, graphics, trademarks, and other intellectual property incorporated into the slide sets remain the sole and exclusive property of the ACP. The slide sets may be used only by the person who downloads or purchases them and only for the purpose of presenting them during not-for-profit educational activities. Users may incorporate the entire slide set or selected individual slides into their own teaching presentations but may not alter the content of the slides in any way or remove the ACP copyright notice. Users may make print copies for use as hand-outs for the audience the user is personally addressing but may not otherwise reproduce or distribute the slides by any means or media, including but not limited to sending them as e-mail attachments, posting them on Internet or Intranet sites, publishing them in meeting proceedings, or making them available for sale or distribution in any unauthorized form, without the express written permission of the ACP. Unauthorized use of the In the Clinic slide sets will constitute copyright infringement.

Toolkit

Want to Subscribe?

Learn more about subscription options

Advertisement
Related Articles
Related Point of Care
Topic Collections
PubMed Articles

Want to Subscribe?

Learn more about subscription options

Forgot your password?
Enter your username and email address. We'll send you a reminder to the email address on record.
(Required)
(Required)