Improved Glucose Control With Weight Loss, Lower Insulin Doses, and No Increased Hypoglycemia With Empagliflozin Added to Titrated Multiple Daily Injections of Insulin in Obese Inadequately Controlled Type 2 Diabetes

Guidelines from the American Diabetes Association and the European Association for the Study of Diabetes recommend the initiation of basal insulin, usually in conjunction with oral antidiabetes agents, in patients with type 2 diabetes mellitus (T2DM) who have marked hyperglycemia when it is unlikely that another drug will be of sufficient additional benefit (1). However, in clinical practice, insulin therapy is often delayed or not optimized. Barriers to the initiation and optimization of insulin therapy include fear of hypoglycemia and weight gain and concerns over the complexity of the treatment regimen (2,3). Approximately 40–50% of patients with T2DM fail to achieve glycemic control with basal insulin plus oral antidiabetes agents after 24 weeks of treatment (4). Further intensification of the insulin regimen to control hyperglycemia is traditionally achieved with progressive additions of prandial insulin up to multiple daily injections of insulin (MDI insulin). However, MDI insulin is overwhelming for most patients, and despite combination with metformin, insulin requirements are often high and patients are still unable to achieve glycemic control. In addition, these patients are typically insulin resistant (5), and comorbidities such as obesity and hypertension are very common. Weight gain associated with the use of high insulin doses may make it even more difficult to achieve glycemic control (5). Accordingly, there is a high unmet need for antidiabetes therapies that can be used in combination with MDI insulin to improve glycemic control in patients with T2DM without exacerbating comorbidities.

Empagliflozin, a potent and selective sodium glucose cotransporter 2 (SGLT2) inhibitor (6), reduces hyperglycemia by reducing renal glucose reabsorption, causing urinary glucose excretion (7). In phase III studies, empagliflozin used as monotherapy or as an add-on to metformin, metformin plus sulfonylurea, pioglitazone (with or without metformin), or basal insulin improved glycemic control and consistently reduced body weight and blood pressure (8–12). Empagliflozin is potentially an attractive option for use in combination with insulin, as its mechanism of action is independent of β-cell function (13), it generates weight loss, and it is associated with a low risk of hypoglycemia (8–12).

The EMPA-REG MDI trial was undertaken to evaluate the efficacy and safety of empagliflozin 10 mg and empagliflozin 25 mg versus placebo as an add-on to MDI insulin with or without metformin for 52 weeks in patients with T2DM and insufficient glycemic control. We hypothesized that empagliflozin would improve glycemic control and body weight in this difficult-to-treat, obese population of patients with T2DM and high insulin requirements.

This was a randomized, double-blind, placebo-controlled, parallel-group study conducted from March 2011 to April 2013 in 104 centers across 14 countries. The clinical trial protocol was approved by the institutional review boards and independent ethics committees and competent authorities of the participating centers, and the study was conducted in accordance with the Declaration of Helsinki and the International Conference on Harmonization Harmonized Tripartite Guideline for Good Clinical Practice. The trial was registered with ClinicalTrials.gov (NCT01306214). All patients provided written informed consent.

This study enrolled obese adults (BMI ≥30 and ≤45 kg/m²) with T2DM and insufficient glycemic control (HbA_1c ≥7.5 to ≤10% [≥58 to ≤86 mmol/mol] at screening) despite diet and exercise counseling and treatment with MDI insulin (total daily dose >60 international units) alone or in combination with metformin (immediate or extended release, ≥1,500 mg/day, maximum tolerated dose, or maximum dose according to the local label). For ≥12 weeks prior to randomization, insulin dose was not to be changed by >10% and metformin dose was to be unchanged. Premixed insulins were not permitted.

Exclusion criteria included uncontrolled hyperglycemia (glucose level >13.3 mmol/L after an overnight fast during the placebo run-in, confirmed by a second measurement); acute coronary syndrome, stroke, or transient ischemic attack within 3 months prior to consent; indication of liver disease; impaired renal function during screening or run-in (estimated glomerular filtration rate [eGFR] using the modification of diet and renal disease equation <60 mL/min/1.73 m²); gastrointestinal surgeries that induce malabsorption; history of cancer (except for basal cell carcinoma) within 5 years; disorders causing hemolysis or unstable erythrocytes; treatment with systemic steroids at time of consent; change in dosage of thyroid hormones within 6 weeks prior to consent; treatment with antiobesity drugs or alcohol or drug abuse within 3 months of consent; and investigational drug intake within 30 days of intake of study drug.

Following a 2-week, open-label, placebo run-in period, patients still meeting the inclusion/exclusion criteria were randomized (1:1:1) to receive once-daily empagliflozin 10 mg, empagliflozin 25 mg, or placebo as an add-on to MDI insulin, with or without metformin, for 52 weeks. Randomization was undertaken using a third-party interactive voice- and web-response system and was stratified by HbA_1c (<8.5, ≥8.5% [<69, ≥69 mmol/mol]), eGFR (chronic kidney disease stage 1, ≥90 mL/min/1.73 m²; chronic kidney disease stage 2, 60–89 mL/min/1.73 m²), region (Europe, North America, Latin America), and background antidiabetes therapy (insulin alone, insulin plus metformin). Patients received diet and exercise counseling based on local recommendations at the start of the run-in period and were reminded to follow their diet and exercise plan at every study visit. For the first 18 weeks, the total daily dose of insulin was to remain within 10% of the prescribed dose at randomization. During the titrated treat-to-target period (weeks 19–40), insulin dose was to be adjusted to achieve a preprandial glucose target of <5.5 mmol/L (<100 mg/dL) and a postprandial glucose target of <7.8 mmol/L (<140 mg/dL). Between weeks 41 and 52, the total daily dose of insulin was to remain within 10% of the insulin dose prescribed at week 40, except for adjustments for safety reasons. Metformin dose was to remain unchanged throughout the study. Study visits were scheduled at screening; the start of the placebo run-in; randomization; after 4, 8, 12, 18, 24, 32, 40, 46, and 52 weeks of treatment; and 4 weeks after the last dose of study drug.

Rescue could be initiated at any time during the treatment period if a patient had hypoglycemia that would put them at risk. During weeks 1–12, rescue medication could be initiated if a patient had a glucose level of >13.3 mmol/L (>240 mg/dL) after an overnight fast. During weeks 13–24, rescue therapy could be initiated if confirmed fasting glucose levels were >11.1 mmol/L (>200 mg/dL). During weeks 24–52, rescue therapy could be initiated if a patient had a confirmed glucose level >10.0 mmol/L (>180 mg/dL) after an overnight fast or HbA_1c >8.0% (>64 mmol/mol) after an overnight fast. The initiation, choice (excluding other SGLT2 inhibitors), and dosage of rescue medication were at the discretion of the investigator according to local prescribing information. Changes in dose of metformin for ≥7 days or addition of a new antidiabetes agent for ≥7 days were considered rescue therapy. Before week 18, changes in total insulin daily dose >10% of the baseline dose for ≥7 days were considered rescue therapy; changes in insulin dose were not considered rescue therapy for the efficacy analyses after week 18.

The primary end point was the change from baseline in HbA_1c at week 18. Secondary end points were changes from baseline at week 52 in insulin daily dose, body weight, and HbA_1c. Exploratory end points included changes from baseline in body weight at week 18; changes from baseline at weeks 18 and 52 in fasting plasma glucose (FPG), systolic blood pressure (SBP), and diastolic blood pressure (DBP); percentage of patients with HbA_1c ≥7% (≥53 mmol/mol) at baseline who had HbA_1c <7% at weeks 18 and 52; and use of rescue therapy.

Safety end points included vital signs, clinical laboratory parameters, and adverse events (AEs) up to 7 days after the last dose of study drug (preferred terms coded according to the Medical Dictionary for Drug Regulatory Activities version 15.1). AEs of special interest included confirmed hypoglycemic AEs (plasma glucose ≤3.9 mmol/L and/or requiring assistance) and events consistent with urinary tract and genital infections identified using prospectively defined search categories based on 73 and 89 preferred terms, respectively.

Sample size calculations indicated that 555 patients (185 per treatment group) would provide 90% power to detect a difference between empagliflozin and placebo in the primary end point with a two-sided significance level of 0.025, assuming a dropout of 15%.

The primary efficacy analysis was performed on the full analysis set (FAS), which included patients treated with ≥1 dose of study drug who had a baseline HbA_1c value. Secondary end points and changes in insulin dose corrected for body weight were analyzed in the “PPS-completers-52” set, defined as patients in the FAS who were on treatment up to day 357 and did not have important protocol violations (such as changes in insulin dose of >10% of prescribed dose at week 40 for ≥7 days during the last 12 weeks). Efficacy analyses of other end points were performed on the FAS at week 18 and in the PPS-completers-52 set at week 52. Safety analyses were performed on the treated set (patients treated with ≥1 dose of study drug).

The primary end point was assessed using an ANCOVA model, with treatment, region, background antidiabetes therapy, and eGFR as fixed effects and baseline HbA_1c as a linear covariate. Secondary end points, continuous exploratory end points, and changes in insulin dose corrected for body weight were analyzed using the statistical model described for the primary end point, with the baseline value for the end point in question as an additional linear covariate. Categorical change in HbA_1c was analyzed using logistic regression including the same factors as covariates.

Values after rescue therapy were set to missing and imputed using the last observation carried forward (LOCF) approach, except for the analysis of lipid parameters, for which LOCF was used and values after rescue therapy were included (LOCF-IR). Changes over time in HbA_1c and insulin daily dose were analyzed using restricted maximum likelihood–based mixed-model repeated measures (MMRM) using observed cases (OCs). Categorical response in HbA_1c was analyzed using a noncompleters considered failures imputation, which assumed that patients who prematurely discontinued the trial and/or received rescue therapy did not achieve the HbA_1c target.

Treatment differences versus placebo for primary and secondary end points were tested using a hierarchical testing approach for each dose using pairwise comparisons between each dose of empagliflozin and placebo using the adjusted means from the model, at a significance level of 2.5% (two-sided) to maintain the overall type I error at 5% in the following sequence: empagliflozin versus placebo in change from baseline in HbA_1c at week 18; empagliflozin versus placebo in change from baseline in insulin dose at week 52; empagliflozin versus placebo in change from baseline in body weight at week 52; empagliflozin versus placebo in change from baseline in HbA_1c at week 52 (one-sided noninferiority test at level of 1.25%), empagliflozin versus placebo in change from baseline in HbA_1c at week 52 (superiority). Exploratory tests were two-sided at a 5% level (no multiplicity adjustment).

A total of 563 patients were randomized and treated with placebo (n = 188), empagliflozin 10 mg (n = 186), or empagliflozin 25 mg (n = 189) (Supplementary Fig. 1). Overall, 475 (84%) patients completed the 52-week treatment period. Baseline mean ± SD characteristics shown in Supplementary Table 1 are age 56.7 ± 9.5 years; BMI 34.8 ± 4.1 kg/m²; FPG 8.52 ± 2.64 mmol/L; HbA_1c 8.3 ± 0.7% [67 ± 7.7 mmol/mol], with 33% of patients with HbA_1c <8% (<64 mmol/mol) and 23% with HbA_1c ≥9% (≥75 mmol/mol); SBP 133.3 ± 15.5 mmHg; DBP 78.8 ± 8.6 mmHg; and baseline insulin dose 92 ± 44 international units. Baseline basal insulin doses were 53 ± 32, 49 ± 23, and 51 ± 25 international units/day with placebo, empagliflozin 10 mg, and empagliflozin 25 mg, respectively. Baseline prandial insulin doses were 40 ± 29, 40 ± 29, and 41 ± 35 international units/day with placebo, empagliflozin 10 mg, and empagliflozin 25 mg, respectively.

During the first 18 weeks of treatment, the total insulin daily dose was to remain within 10% of the prescribed dose at randomization. At week 18, adjusted mean ± SE changes from baseline in HbA_1c were −0.50 ± 0.05% (−5.5 ± 0.5 mmol/mol) with placebo compared with −0.94 ± 0.05% (−10.3 ± 0.5 mmol/mol) with empagliflozin 10 mg and −1.02 ± 0.05% (−11.1 ± 0.5 mmol/mol) with empagliflozin 25 mg (P < 0.001 for both) (Table 1, Fig. 1A).

Adjusted mean ± SE changes from baseline in FPG were 0.19 ± 0.16 mmol/L with placebo compared with −0.98 ± 0.16 mmol/L with empagliflozin 10 mg and −1.36 ± 0.16 mmol/L with empagliflozin 25 mg at week 18 (P < 0.001 vs. placebo for both) (Table 1, Supplementary Fig. 2A).

Body weight increased from baseline with placebo (adjusted mean ± SE 0.34 ± 0.18 kg) compared with a decrease with empagliflozin 10 mg (−0.97 ± 0.18 kg) and empagliflozin 25 mg (−1.54 ± 0.18 kg) at week 18 (P < 0.001 vs. placebo for both) (Table 1, Fig. 1B).

Adjusted mean ± SE changes from baseline in SBP were −1.2 ± 0.8 mmHg with placebo compared with −3.6 ± 0.8 mmHg with empagliflozin 10 mg (P = 0.037). The change from baseline in SBP with empagliflozin 25 mg did not reach significance versus placebo (Supplementary Table 2, Supplementary Fig. 3A). Adjusted mean ± SE changes from baseline in DBP did not reach significance for either empagliflozin dose (Supplementary Table 2, Supplementary Fig. 3B).

During the first 18 weeks, 15 patients (8.0%) on placebo, 3 patients (1.6%) on empagliflozin 10 mg, and 5 patients (2.6%) on empagliflozin 25 mg received rescue therapy. The rescue therapy used most often was an increase in insulin dose.

Insulin dose was to be adjusted during weeks 19–40 to reach glucose targets and then to remain within 10% of the prescribed dose at week 40 during weeks 41–52. Adjusted mean daily dose of insulin over time is shown in Fig. 2A. Adjusted mean HbA_1c levels over 52 weeks are shown in Fig. 2B. At week 52, adjusted mean ± SE changes from baseline were 10.2 ± 2.2 international units/day with placebo compared with 1.3 ± 2.1 international units/day with empagliflozin 10 mg and −1.1 ± 2.1 international units/day with empagliflozin 25 mg (P = 0.004 for empagliflozin 10 mg; P < 0.001 for empagliflozin 25 mg) (Table 1, Fig. 2C). Corrected for body weight, adjusted mean ± SE changes from baseline were 0.10 ± 0.02 international units/kg with placebo compared with 0.03 ± 0.02 international units/kg with empagliflozin 10 mg (difference of adjusted means vs. placebo, −0.07 international units/kg [95% CI −0.12 to −0.01]; P = 0.023) and 0.00 ± 0.02 international units/kg with empagliflozin 25 mg (difference of adjusted means vs. placebo, −0.10 international units/kg [95% CI −0.15 to −0.04]; P < 0.001). Mean ± SE changes from baseline in basal insulin dose at week 52 were 9.2 ± 1.9 international units/day for placebo, 3.7 ± 1.5 international units/day for empagliflozin 10 mg, and 2.5 ± 1.5 international units/day for empagliflozin 25 mg. For prandial insulin dose, mean ± SE changes from baseline at week 52 were 0.3 ± 1.3 international units/day for placebo, −1.9 ± 1.0 international units/day for empagliflozin 10 mg, and −3.5 ± 1.3 international units/day for empagliflozin 25 mg. There were no changes from baseline in the number of prandial insulin shots per day in any treatment group.

At week 52, due to the insulin titration in the placebo group, the adjusted mean ± SE change from baseline in HbA_1c was −0.81 ± 0.08% (−8.9 ± 0.9 mmol/mol). With much lower insulin titration in the empagliflozin groups, changes were −1.18 ± 0.08% (−12.9 ± 0.9 mmol/mol) with empagliflozin 10 mg and −1.27 ± 0.08% (−13.9 ± 0.9 mmol/mol) with empagliflozin 25 mg (P < 0.001 for both) (Table 1, Fig. 2D). Of note, a significant proportion of patients with HbA_1c ≥7.0% (≥53 mmol/mol) at baseline reached HbA_1c <7.0% with empagliflozin 10 mg (31.4%) and empagliflozin 25 mg (41.7%; P < 0.01 for odds ratio vs. placebo [21.0%]) (Fig. 2E).

Adjusted mean ± SE changes from baseline in FPG at week 52 were −0.63 ± 0.19 mmol/L with placebo compared with −1.32 ± 0.19 mmol/L with empagliflozin 10 mg (P = 0.012) and −1.43 ± 0.19 mmol/L with empagliflozin 25 mg (P = 0.004) (Table 1, Supplementary Fig. 2B).

Treatment with empagliflozin significantly reduced body weight at week 52. Adjusted mean ± SE changes from baseline were 0.44 ± 0.36 kg with placebo compared with −1.95 ± 0.36 kg with empagliflozin 10 mg and −2.04 ± 0.36 kg with empagliflozin 25 mg (differences of adjusted means vs. placebo were −2.39 kg [95% CI −3.40 to −1.39] for empagliflozin 10 mg and −2.48 kg [95% CI −3.48 to −1.47] for empagliflozin 25 mg; P < 0.001 for both) (Table 1, Fig. 2F).

Changes from baseline in SBP with both doses of empagliflozin and in DBP with empagliflozin 10 mg did not reach significance versus placebo at week 52 (Supplementary Table 2, Supplementary Fig. 3C and D). Mean ± SE change from baseline in DBP was greater with empagliflozin 25 mg than placebo (−2.5 ± 0.6 vs. −0.5 ± 0.6 mmHg; P = 0.035) at week 52 (Supplementary Table 2, Supplementary Fig. 3D). No clinically meaningful changes in pulse rate were noted.

Data on AEs are shown in Table 2. Over the 52-week treatment period, the proportions of patients with ≥1 AE, ≥1 serious AE, or ≥1 AE leading to discontinuation were similar across the treatment groups, with ∼90% of patients reporting only mild or moderate events. One death occurred (metastatic lung cancer in a patient in the empagliflozin 25 mg group).

The proportion of patients with confirmed hypoglycemic AEs up to week 18 was slightly higher in the empagliflozin 10 mg group (74 patients [39.8%]) and the empagliflozin 25 mg group (78 patients [41.3%]) compared with the placebo group (70 patients [37.2%]). However, only one patient in each of the three treatment groups had severe hypoglycemic AEs (requiring assistance). Over the full 52-week treatment period, including the treat-to-target period, the proportion of patients with confirmed hypoglycemic AEs was similar in all treatment groups (placebo, 109 patients [58.0%]; empagliflozin 10 mg, 95 patients [51.1%]; empagliflozin 25 mg, 109 patients [57.7%]). Three patients each in the placebo and empagliflozin 10 mg groups and one in the empagliflozin 25 mg group had severe hypoglycemia (requiring assistance).

Over the 52-week treatment period, events consistent with urinary tract infection were reported in similar proportions of patients on placebo (15.4%), empagliflozin 10 mg (15.6%), and empagliflozin 25 mg (15.3%). More female patients (24.8–27.0%) than male patients (0–5.2%) reported these events across all treatment groups. The majority of patients who reported any events consistent with urinary tract infection reported only mild events. One patient (on empagliflozin 25 mg) reported a severe event, one patient (on empagliflozin 25 mg) reported ≥1 event that required hospitalization, and no urinary tract infection events led to study discontinuation. Events consistent with genital infections were reported in more patients in the empagliflozin 10 mg (4.3%) and empagliflozin 25 mg (9.5%) groups compared with the placebo group (1.6%). These events were reported in a greater proportion of female patients on empagliflozin 10 mg (7/89 [7.9%]) or empagliflozin 25 mg (11/105 [10.5%]) than on placebo (2/118 [1.8%]) and in a greater proportion of male patients on empagliflozin 25 mg (7/84 [8.3%]) than on empagliflozin 10 mg (1/97 [1.0%]) or placebo (1/75 [1.3%]). All such events were mild or moderate in intensity, only one patient (on empagliflozin 25 mg) experienced an event that led to discontinuation, and no patients reported events that required or prolonged hospitalization.

Changes in laboratory parameters are shown in Supplementary Table 3. There were small decreases from baseline in uric acid and small increases from baseline in hematocrit (∼4%) with empagliflozin. Electrolyte levels were unchanged across treatment groups. There were small decreases from baseline to end of treatment in mean eGFR in all treatment groups, which returned to baseline levels at follow-up in the empagliflozin groups. No major differences between placebo and empagliflozin in mean changes from baseline in total cholesterol, LDL cholesterol, HDL cholesterol, or triglycerides were noted.

This is the first dedicated trial to determine the efficacy and safety of an SGLT2 inhibitor in the difficult-to-treat population of obese, insulin-resistant patients with insufficient glycemic control despite high-dose MDI insulin. Compared with insulin titrations alone, which resulted in HbA_1c reductions to 7.5% (58 mmol/mol), the addition of empagliflozin 10 mg and empagliflozin 25 mg to titrated MDI insulin improved glycemic control further, despite lower insulin doses, to achieve HbA_1c levels of 7.2 and 7.1% (55 and 54 mmol/mol), respectively, with no increase in the risk of hypoglycemia and with weight loss.

The use of insulin in patients with T2DM is associated with a high risk of hypoglycemic events and weight gain, complicating the management of hyperglycemia (14). Fear of hypoglycemia and weight gain decrease adherence to medication (15,16), while insulin-induced weight gain in patients with T2DM is associated with worsening of insulin resistance (1), resulting in a vicious cycle. This study investigating the effect of empagliflozin as an add-on to MDI insulin is unique in its study design, as it comprised three distinct insulin dosing periods. In a stable MDI insulin period during the first 18 weeks, empagliflozin resulted in significant HbA_1c reduction and weight loss compared with placebo, with a slightly increased frequency of mild hypoglycemic events but no increase in severe hypoglycemic events. An increased frequency of hypoglycemia has previously been reported when antidiabetes agents with a low risk of hypoglycemia are added to insulin (17,18). Of note, over the full 52-week treatment period, including the treat-to-target period, the proportion of patients with confirmed hypoglycemic AEs was similar in all treatment groups. This might be explained by differences in insulin titration in the empagliflozin and placebo arms due to the insulin-independent mechanism of action of empagliflozin (13); the incomplete inhibition of renal glucose reabsorption by empagliflozin (19); a diminished effect of SGLT2 inhibition at low glucose levels due to physiological decline in glomerular filtration rate (due to sympathetic nervous system activation) (20); a compensatory increase in hepatic gluconeogenesis (21); or a combination of these factors.

The change from baseline in HbA_1c in the placebo group at week 18 was more pronounced than previously observed in studies with empagliflozin. We may speculate that changes in diet and lifestyle due to participation in a dedicated MDI insulin trial may itself have resulted in improved glycemic control in this obese, insulin-resistant population, reinforcing the general paradigm that pharmacological treatment of diabetes has to be accompanied by diet and lifestyle intervention.

Empagliflozin was well tolerated when used in combination with MDI insulin with or without metformin for 52 weeks. Empagliflozin was not associated with a higher rate of events consistent with urinary tract infections but was associated with an increased risk of events consistent with genital infection, as has been observed in other studies of SGLT2 inhibitors (22). However, no events consistent with genital infection were severe, and such events led to premature discontinuation in only one patient. Small decreases in eGFR during treatment with empagliflozin likely reflected hemodynamic changes, as eGFR had returned to baseline levels 4 weeks after treatment discontinuation, consistent with the results of another phase III study that examined the effect of empagliflozin on renal function over time (23).

Limitations of this trial include that insulin titration was at the investigator’s discretion based on prespecified treatment goals and was not defined or enforced by an independent insulin titration monitoring committee. Therefore, the full effect of empagliflozin on titrated insulin doses cannot be fully assessed. Furthermore, the conclusions from this study are limited to the population studied and are not applicable to the general population of patients with T2DM.

In conclusion, in obese patients with T2DM and inadequate glycemic control despite MDI insulin, empagliflozin 10 mg and empagliflozin 25 mg once daily for 52 weeks improved glycemic control and reduced body weight without increasing the risk of hypoglycemia and with lower insulin requirements. This suggests that empagliflozin may provide a new treatment option for this challenging-to-treat patient population.

Acknowledgments. Medical writing assistance was provided by Clare Ryles and Wendy Morris of Fleishman-Hillard Group Ltd., during the preparation of this article.

Duality of Interest. This study was funded by Boehringer Ingelheim and Eli Lilly. A.J., G.F., A.S., G.K., H.J.W., and U.C.B. are employees of Boehringer Ingelheim. J.R. has served on scientific advisory boards and received honoraria or consulting fees from companies involved in development of SGLT2 inhibitors, including Bristol-Myers Squibb, Roche, GlaxoSmithKline, Johnson & Johnson, Boehringer Ingelheim, and Lexicon, and has also received grants/research support from Pfizer, Roche, Bristol-Myers Squibb, GlaxoSmithKline, AstraZeneca, Johnson & Johnson, Boehringer Ingelheim, and Lexicon. Medical writing assistance was supported financially by Boehringer Ingelheim. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. J.R. contributed to the acquisition and interpretation of data and drafted the manuscript. A.J., G.F., A.S., and G.K. contributed to the interpretation of data and reviewed/edited the manuscript. H.J.W. and U.C.B. contributed to the study design and interpretation of data and reviewed/edited the manuscript. The authors are fully responsible for all content and editorial decisions, were involved at all stages of manuscript development, and have approved the final version. G.F. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented at the 74th Scientific Sessions of the American Diabetes Association, San Francisco, CA, 13–17 June 2014.