Clinical Parameters, Fuel Oxidation, and Glucose Kinetics in Patients With Type 2 Diabetes Treated With Dapagliflozin Plus Saxagliptin

Despite irrefutable evidence that improved glycemic control prevents/delays the onset of microvascular complications (1,2), it has been reported that ∼50% of patients with type 2 diabetes fail to achieve the American Diabetes Association HbA_1c goal of <7.0% (53 mmol/mol) for glycemic control (3). Progressive β-cell failure, weight gain, and hypoglycemia are major obstacles for achieving optimal glycemic control (4,5). Therefore, treatment regimens that effectively lower plasma glucose, maintain durable glycemic control, do not cause hypoglycemia, and are not associated with weight gain are needed. Dipeptidyl peptidase 4 inhibitors (DPP4i) and renal sodium–glucose cotransporter 2 inhibitors (SGLT2i) effectively reduce HbA_1c without causing hypoglycemia (6,7), and the decrease in HbA_1c persists for >4 years (8). DPP4i reduce plasma glucose by stimulating insulin and suppressing glucagon secretion, leading to a decrease in endogenous glucose production (EGP) (9). SGLT2i as monotherapy or in combination with other antidiabetic agents reduce HbA_1c by 0.8–1.0% (8,10,11). The primary action of SGLT2i is on the kidney to promote glucosuria (12–14), and the subsequent decline in plasma glucose secondarily results in improved β-cell function and insulin sensitivity (i.e., reversal of glucotoxicity) (15–17).

The primary determinant of fasting plasma glucose (FPG) concentration is basal rate of EGP, which primarily is derived from liver (18). When SGLT2i are given to subjects with normal glucose tolerance, FPG does not change despite marked glucosuria because basal EGP increases to precisely offset the increased rate of renal glucose excretion (19). When SGLT2i are given to patients with type 2 diabetes, the increase in EGP begins when plasma glucose is well within the hyperglycemic range and offsets by ∼50% the amount of glucose excreted in urine (14,15,19). In these studies, SGLT2i therapy was associated with an increase in plasma glucagon, a decrease in plasma insulin, and an increase in insulin/glucagon (I/G) ratio. We hypothesized that the addition of an agent that stimulates insulin and inhibits glucagon would prevent an increase in EGP in SGLT2i-treated patients and produce a greater decrease in FPG and HbA_1c. Because DPP4i agents stimulate insulin and inhibit glucagon secretion (6,9), we envisioned that combined DPP4i/SGLT2i therapy would prevent the compensatory increase in EGP and enhance the decrease in FPG and HbA_1c. Our hypothesis is partly based on results reported by Hansen et al. (20), whereby treatment with saxagliptin (DPP4i) plus dapagliflozin (SGLT2i) was conducted in patients with poorly controlled type 2 diabetes. This study produced only a very modest decrease in HbA_1c compared with each agent alone, and the mechanism behind the changes was not entirely elucidated. In the current study, we tested the hypothesis that the addition of DPP4i to SGLT2i would provide superior glucose-lowering efficacy compared with SGLT2i monotherapy by inhibiting the rise in EGP that occurs in response to SGLT2i-induced glucosuria.

Fifty-six subjects with poorly controlled (HbA_1c 7.0–11.0% [53–97 mmol/mol]) type 2 diabetes were randomized to dapagliflozin 10 mg/day plus placebo (DAPA), DAPA 10 mg/day plus saxagliptin 5 mg/day (DAPA/SAXA), or placebo plus placebo (PCB). Other than diabetes, subjects were in good health as determined by medical history, physical examination, routine blood chemistry analyses, urinalysis, and electrocardiogram. Patients with type 2 diabetes were controlled by diet (n = 4) or a stable dose (>3 months) of sulfonylurea (n = 1), metformin (n = 18), or both (n = 33). Body weight was stable (±3 pounds) in all subjects for at least 3 months before study, and no subject participated in an excessively heavy exercise program. Patients were randomized (1:2:2 ratio) to receive PCB, DAPA, or DAPA/SAXA acutely and for 16 weeks in a double-blinded fashion. Twelve patients (10 male, 2 female) entered the PCB group with a mean age of 50 ± 2 years and diabetes duration of 6.8 ± 1.5 years. Twenty-two patients (8 male, 14 female, mean age 52 ± 2 years, diabetes duration of 7.0 ± 1.4 years) were randomized into the DAPA group and 22 into the DAPA/SAXA group (9 male, 13 female, mean age 52 ± 2 years, diabetes duration of 7.1 ± 1.3 years). Additional subject characteristics at baseline and following 16 weeks of therapy are summarized in Table 1. Studies were carried out in the clinical research center (CRC) of the Texas Diabetes Institute–University Health System. Protocols were approved by the UT Health San Antonio institutional review board, and informed written consent was obtained from each patient before participation.

At the screening visit, patients arrived at CRC at 0800 h after a 10- to 12-h overnight fast for determination of body composition with DEXA (Discovery Wi; Hologic, Bedford, MA). On a subsequent day (day 1 [Pre-TX]), participants returned to CRC at 0600 h after a 10-h overnight fast for measurement of EGP. A catheter was placed into an antecubital vein, and prime (25 μCi × FPG/100)-continuous (0.25 μCi/min) infusion of 3-³H-glucose (DuPont, Willington, DE) was started and continued until 1500 h. At 0800 h, a second catheter was placed retrogradely into a vein on the dorsum of the hand, which was placed in a heated box (70°C) for sampling of arterialized blood. Indirect calorimetry was performed for 45 min using a ventilated hood system (Vmax 29n; Sensormedics, Yorba Linda, CA) to determine rates of glucose and lipid oxidation. After 150 min of tracer equilibration, arterialized blood samples were drawn at −30, −20, −15, −10, −5, and 0 for determination of plasma glucose, insulin, C-peptide, glucagon, free fatty acid (FFA), and β-OH-butyrate (BOHB) concentrations and tritiated glucose–specific activity. At time 0 (0900 h), patients were randomized (1:2:2 ratio) to receive PCP, DAPA, or DAPA/SAXA as outlined above. After time 0, plasma samples were obtained every 15 min for 300 min for determinations as described above. Urine was collected from 0600 to 0900 h and from 0900 to 1400 h for measurement of urinary glucose excretion (UGE) rate. At 1400 h, infusions were discontinued, and subjects received a meal and were discharged. All data obtained on day 1 were considered baseline data (i.e., before exposure to drug[s] [Pre-Tx]). In addition, on day 1 (Pre-Tx), baseline glucose kinetics were determined during a 300-min period following acute exposure to drug(s). Participants were instructed to take the medication to which they were assigned every morning for 16 weeks and continue their background medication. Patients returned every 4 weeks for follow-up, including vital signs, weight, brief physical examination, and FPG and HbA_1c measurement. During the 16-week study, one patient in the PCB group required rescue therapy (HbA_1c >11.0% [97 mmol/mol]) and was excluded from analysis. At week 16 (Post-Tx), all baseline measurements, including the determination of glucose kinetics after 300 min of exposure to drug(s), were repeated, and patients were instructed to return to their primary care provider.

Plasma glucose was measured with glucose oxidase method (Analox Technologies, Toronto, Ontario, Canada). Plasma insulin (IBL America, Minneapolis, MN) and C-peptide (MP Biomedicals, Santa Ana, CA) were determined with immunoradiographic assays and plasma glucagon (MilliporeSigma, Burlington, MA) by radioimmunoassay. Fluorometric assays were used to measure plasma BOHB (intra-assay coefficient of variance [CV] 4.05%, interassay CV 3.18%; Cayman Chemical, Ann Arbor, MI) and FFA (intra-assay CV 4.91%, interassay CV 0.75%; FUJIFILM Wako Diagnostics, Richmond, VA).

One of the primary end points was change in EGP from baseline (−30 to 0 min) to a 240- to 300-min period following acute administration of DAPA and DAPA/SAXA compared with PCB. EGP before drug administration also was compared with EGP after 16 weeks of therapy between groups. Under steady-state postabsorptive conditions, the basal R_a equals 3-³H-glucose infusion rate divided by steady-state plasma tritiated glucose–specific activity. After drug administration, nonsteady conditions for 3-³H-glucose–specific activity (tracer/tracee ratio) prevail, and rates of total body R_a and R_d are calculated using the Steele equation (21). Tissue R_d is derived by subtracting the UGE rate from total R_d. Glucose metabolic clearance rate (MCR_glu) was calculated by dividing total R_d by plasma glucose concentration. Substrate oxidation rates were calculated using standard equations for VO₂ and VCO₂ (22,23).

Within each group, change from baseline (before drug ingestion on day 0) in EGP, total R_d, tissue R_d, UGE, substrate oxidation rates, and plasma metabolite/hormone levels was compared with each variable after drug administration using paired Student t test. Between-group comparisons (PCB vs. DAPA, PCB vs. DAPA/SAXA, and DAPA vs. DAPA/SAXA) for the change from baseline to study end for all preceding variables was compared using ANCOVA. Post hoc testing was done with Bonferroni correction. After 16 weeks, change in variables compared with day 0 (before drug administration) within each group was performed with Student t test, while between-group comparisons after 16 weeks were assessed using ANCOVA with Bonferroni correction. Two-sided α ≤ 0.05 was considered to be statistically significant.

Power calculations indicated that 21 patients per treatment group and 11 patients in the PCB group would provide 90% power to detect a difference in EGP of 0.28 mg/kg/min among treatment groups with a 95% CI of 0.17 mg/kg/min. This calculation used a within-subject SD estimate of 0.24 mg/kg/min with α = 0.05 and two-sided test on the basis of previous results from our research center (24).

The decrease in HbA_1c was significantly greater in DAPA/SAXA versus DAPA (−2.0 ± 0.3 vs. −1.4 ± 0.2%, P < 0.01), and both were greater (P < 0.001) than PCB (0.1 ± 0.3%). The decrease in FPG was greater in DAPA/SAXA versus DAPA (−77 ± 10 vs. −47 ± 8 mg/dL, P < 0.01), and both were greater than PCB (1 ± 14 mg/dL, P < 0.001). Body weight and BMI decreased similarly in DAPA and DAPA/SAXA and significantly more than PCB, although the decrease in percent body weight was similar in all three groups (Table 1).

Following placebo ingestion, FPG decreased from 182 ± 6 to 153 ± 8 mg/dL during the 300-min study period, most likely because of fasting. After the administration of DAPA (192 ± 5 to 130 ± 8 mg/dL) and DAPA/SAXA (196 ± 6 to 141 ± 7 mg/dL), the decline in plasma glucose was significantly greater (P < 0.01) than PCB (DAPA vs. DAPA/SAXA, P not significant) (Fig. 1A). Baseline EGP was similar in all three groups and decreased from 2.40 ± 0.16 to 1.96 ± 0.12 mg/kg/min (P < 0.05) after PCB, whereas in DAPA (2.47 ± 0.12 to 2.26 ± 0.11 mg/kg/min) and DAPA/SAXA (2.48 ± 0.12 to 2.27 ± 0.10 mg/kg/min), the decline in EGP was significantly less (P < 0.01) (Fig. 1B).

Plasma insulin, C-peptide, and glucagon concentrations did not change significantly in PCB (Table 2 and Fig. 2A–C). Plasma insulin declined in DAPA (P < 0.05) and remained unchanged in DAPA/SAXA. The decrease in plasma C-peptide (4.6 ± 0.7 to 4.1 ± 0.7 ng/mL) was not significant and, plasma glucagon did not change in either group during the 300-min study period (Table 2 and Fig. 2A–C). Following DAPA and DAPA/SAXA, UGE increased similarly and markedly (P < 0.001) (Table 2). Total R_d was elevated in DAPA and DAPA/SAXA, reflecting the increase in UGE, but tissue R_d decreased (Table 2), paralleling the decline in plasma glucose concentration (Fig. 1A). Baseline MCR_glu did not change during the 300-min study period in PCB (1.45 ± 0.09 vs. 1.51 ± 0.10 mL/kg/min), whereas in DAPA (1.52 ± 0.08 to 2.02 ± 0.12 mL/kg/min) and DAPA/SAXA (1.46 ± 0.08 to 1.85 ± 0.08 mL/kg/min), significant increases in MCR_glu were detected after acute exposure to the drugs (P < 0.01) (Table 2). Considering that the glucose excreted in urine is not metabolized, MCR_glu calculated using tissue R_d (instead of total R_d) shows that in the Pre-Tx study, tissue MCR_glu did not change in the PCB group (1.43 ± 0.09 vs. 1.51 ± 0.08 mL/kg/min), the DAPA group (1.49 ± 0.09 vs. 1.54 ± 0.10 mL/kg/min), or the DAPA/SAXA group (1.45 ± 0.08 vs. 1.43 ± 0.07 mL/kg/min) during the 300-min study period.

Following 16 weeks of chronic treatment with DAPA and DAPA/SAXA, FPG declined by −47 ± 8 and −77 ± 10 mg/dL, respectively, and did not change in PCB (Table 1 and Fig. 1C). In the 300-min study period, FPG declined from 190 ± 12 to 149 ± 15 mg/dL (P < 0.05) in PCB and from 145 ± 10 to 114 ± 11 mg/dL (P < 0.05) in DAPA, whereas in the combined DAPA/SAXA group, there was no significant change in FPG (119 ± 7 to 113 ± 7 mg/dL, P not significant) (Fig. 1C).

At week 16, basal EGP with PCB (2.47 ± 0.18 mg/kg/min) remained unchanged versus day 1 (2.40 ± 0.20 mg/kg/min). In contrast, at week 16, basal EGP in both DAPA (2.47 ± 0.13 to 2.67 ± 0.09 mg/kg/min) and DAPA/SAXA (2.48 ± 0.10 to 2.61 ± 0.08 mg/kg/min) increased modestly (P < 0.05), and both were significantly (P < 0.05) greater than PCB. In the 300-min study period, the decline in EGP in PCB (Fig. 1D) was superimposable to that on day 1 of PCB (Fig. 1B). Despite higher baseline values at week 16, the decrement in EGP in both DAPA (0.20 ± 0.12 mg/kg/min) and DAPA/SAXA (0.13 ± 0.14 mg/kg/min) was significantly blunted (P < 0.01) vs. PCB (0.65 ± 0.14 mg/kg/min) during the 300-min study period.

At week 16 in PCB, there were no changes in UGE, and total R_d and tissue R_d decreased slightly (Table 2), paralleling the decline in plasma glucose concentration (Fig. 1C). Following both DAPA and DAPA/SAXA, baseline UGE (which was comparable to UGE after DAPA administration on day 1) did not change significantly (Table 2). Total R_d and tissue R_d increased slightly but not significantly following both DAPA and DAPA/SAXA at week 16 (Table 2). Not surprisingly, therefore, plasma glucose concentration either did not change significantly (DAPA/SAXA) or decreased only slightly (DAPA) (Fig. 1C). At week 16, even though baseline MCR_glu did not change significantly in PCB, in DAPA, it increased by ∼36% to 2.07 ± 0.14 mL/kg/min and further in DAPA/SAXA by ∼46% to 2.13 ± 0.11 mL/kg/min (P < 0.05, DAPA vs. DAPA/SAXA). During the 300-min study period, there was no change in MCR_glu in PCB (1.61 ± 0.16 vs. 1.69 ± 0.14 mL/kg/min), whereas in DAPA (2.07 ± 0.14 vs. 2.46 ± 0.12 mL/kg/min) and DAPA/SAXA (2.13 ± 0.11 vs. 2.38 ± 0.12 mL/kg/min), there were additional increases in MCR_glu (both P < 0.05 vs. PCB) (Table 2). Again, considering that the glucose excreted in the urine is not metabolized, MCR_glu calculated using tissue R_d (instead of total R_d) showed that at week 16, baseline tissue R_d-MCR_glu increased by ∼8% in PCB to 1.61 ± 0.11 mL/kg/min, by ∼10% in DAPA to 1.62 ± 0.06 mL/kg/min, and further by ∼20% in DAPA/SAXA to 1.73 ± 0.10 mL/kg/min (P < 0.05, DAPA vs. DAPA/SAXA). During the 300-min study period, there was no change in tissue R_d-MCR_glu in PCB (1.61 ± 0.14 vs. 1.69 ± 0.12 mL/kg/min), whereas the additional increase in DAPA (1.62 ± 0.06 vs. 1.93 ± 0.08 mL/kg/min) was twofold higher (Δ 0.31 vs. 0.17 mL/kg/min) than in DAPA/SAXA (1.73 ± 0.10 vs. 1.90 ± 0.12 mL/kg/min) (P < 0.05, DAPA vs. DAPA/SAXA).

At week 16, fasting plasma insulin in PCB was unchanged compared with day 1. In both DAPA and DAPA/SAXA, fasting plasma insulin decreased significantly (P < 0.05) compared with PCB (Fig. 2A), paralleling the decline in FPG concentration. Fasting plasma C-peptide concentrations at week 16 did not change and were similar to those at day 1 in PCB, DAPA, and DAPA/SAXA. In all three groups, plasma C-peptide did not change significantly after drug administration (Fig. 2B). Fasting plasma glucagon at week 16 was unchanged from time 0 in the PCB and DAPA groups and, as expected, decreased significantly in the DAPA/SAXA group (Fig. 2C). Plasma glucagon was unchanged from baseline following drug administration in all three groups. On day 1, the baseline I/G ratio with PCB was 0.32 ± 0.04 μU/mL per pg/mL and did not change during the 300-min study period. Following DAPA, there was a small nonsignificant decrease in I/G ratio from 0.33 ± 0.04 to 0.26 ± 0.03 μU/mL per pg/mL, while the I/G ratio did not change following DAPA/SAXA (0.34 ± 0.05 vs. 0.32 ± 0.04 μU/mL per pg/mL). After 16 weeks, the baseline I/G ratio decreased significantly in the DAPA group from 0.33 ± 0.04 to 0.23 ± 0.02 μU/mL per pg/mL (P < 0.05) and did not change in the DAPA/SAXA group (0.34 ± 0.05 vs. 0.33 ± 0.05 μU/mL per pg/mL). At week 16, there were no changes in the I/G ratio during the 300-min study period after drug administration in any of the three groups. On day 1, the baseline C-peptide/glucagon (C-pep/G) ratio in PCB was 0.07 ± 0.01 ng/mL per pg/mL and did not change during the 300-min study period. Following DAPA, there was a small, nonsignificant decrease in C-pep/G ratio from 0.09 ± 0.01 to 0.07 ± 0.02 ng/mL per pg/mL, while the C-pep/G ratio did not change following DAPA/SAXA (0.08 ± 0.02 vs. 0.08 ± 0.02 ng/mL per pg/mL). After 16 weeks, the baseline C-pep/G ratio showed a nonsignificant decrease in the DAPA group (0.09 ± 0.01 vs. 0.07 ± 0.02 ng/mL per pg/mL) and did not change in the DAPA/SAXA group (0.08 ± 0.02 vs. 0.09 ± 0.02 ng/mL per pg/mL). At week 16, there were no changes in the C-pep/G ratio during the 300-min study period after drug administration in any of the three groups.

At week 16, the baseline plasma FFA concentration decreased in PCB (590 ± 50 to 480 ± 40 μmol/L) but increased in DAPA from 570 ± 50 to 610 ± 30 μmol/L (P < 0.05, PCB vs. DAPA). In the DAPA/SAXA group, fasting plasma FFA concentration decreased significantly from 690 ± 50 vs. 600 ± 30 μmol/L (P < 0.05 vs. DAPA). Baseline plasma BOHB levels (∼0.22 mmol/L) did not change significantly in any group at week 16 (range ∼0.25–0.30 mmol/L). In PCB, the basal (following overnight fast) rates of carbohydrate (1.15 ± 0.04 mg/kg/min) and lipid oxidation (0.63 ± 0.01 mg/kg/min) did not change after 16 weeks of therapy (1.10 ± 0.04 and 0.63 ± 0.01 mg/kg/min, respectively). In DAPA, basal rates of carbohydrate oxidation decreased significantly after 16 weeks from 1.13 ± 0.02 to 0.68 ± 0.01 mg/kg/min, and lipid oxidation increased from 0.56 ± 0.01 to 0.75 ± 0.01 mg/kg/min (P < 0.01, Pre-Tx vs. Post-Tx). DAPA/SAXA combination treatment for 16 weeks had no significant effect on basal rates of carbohydrate (1.10 ± 0.02 vs. 0.90 ± 0.02 mg/kg/min) and lipid oxidation (0.60 ± 0.01 vs. 0.63 ± 0.01 mg/kg/min) (Fig. 3).

No serious adverse events were observed in any of the three treatment groups. Three patients randomized to DAPA and two patients randomized to DAPA/SAXA developed a genital mycotic infection, which resolved with topical and/or oral antimycotic treatment. No hypoglycemia, urinary tract infections, or other side effects were observed. Furthermore, there were no adverse effects related to the use of radioisotopes.

In this study, we demonstrate that in patients with type 2 diabetes, combination treatment with DAPA/SAXA for 16 weeks leads to a greater improvement in glycemic control than does DAPA monotherapy. The decline in HbA_1c with DAPA/SAXA was significantly greater than with DAPA alone (−2.0 vs. −1.4%, P < 0.001), and this primarily was accounted for by the greater decline in FPG concentration (77 vs. 47 mg/dL, P < 0.001). These differences could not be accounted for by differences in weight loss or body fat distribution (Table 1) and are in agreement with previously reported findings (20,25). Consistent with results published earlier from our laboratory (13–15,19) and others (17), DAPA, whether administered acutely (day 1) or chronically (week 16) as monotherapy or in combination with SAXA, caused a “compensatory” stimulation of EGP. Nonetheless, despite consistent and equivalent increases in EGP in the DAPA and DAPA/SAXA groups, dual therapy reduced FPG and HbA_1c significantly more than DAPA monotherapy. Glucose kinetic data showed comparable increases in total R_d in both groups, which was entirely explained by the DAPA-induced glucosuria. In absolute terms, tissue R_d decreased significantly in both groups in parallel to the decline in plasma glucose concentration that resulted from increased UGE. Of interest, however, MCR_glu, whether calculated using total or tissue R_d, which reflects the intrinsic ability of tissues to take up glucose, was significantly higher in DAPA/SAXA versus DAPA, with differences of 0.67 ± 0.05 vs. 0.55 ± 0.04 mL/kg/min (0.28 ± 0.03 vs. 0.13 ± 0.04 mL/kg/min if tissue R_d is used) (P < 0.05). The greater elevation in MCR_glu in the DAPA/SAXA group could be explained by changes in plasma FFA concentration and lipid oxidation, both of which increased in DAPA but declined in DAPA/SAXA. In the current study, glucose kinetics was measured at baseline (day 1) and again at week 16. Thus, it is possible that MCR_glu increased to an even greater extent during the 4-month treatment interval, declining with time as plasma glucose concentration declined. It is also possible, although not examined in this study, that improved postprandial glucose metabolism contributed to the greater HbA_1c reduction with DAPA/SAXA in combination versus DAPA monotherapy.

Plasma insulin and C-peptide both decreased slightly after DAPA and remained unchanged following DAPA/SAXA. Since EGP was similar in the DAPA and DAPA/SAXA study groups and failed to decline (as was observed following the PCB administration), the compensatory stimulation of EGP following DAPA could not be explained by the decrease in plasma insulin concentration. Plasma glucagon concentration did not change from baseline in any of the three treatment groups (Table 2), and thus, changes in plasma glucagon cannot explain the compensatory increase in EGP following DAPA either. These hormonal changes are consistent with our previous findings in comparable circumstances (26).

DAPA markedly blunted the decline in EGP observed following PCB on day 1 (Fig. 1B) and after 16 weeks of treatment (Fig. 1D). A compensatory rise in EGP in response to SGLT2i-induced glucosuria has been reported in both short-term and longer-term studies (13–17,26) and has been attributed to a decrease in plasma insulin (C-peptide) concentration and concomitant increase in plasma glucagon concentration, which presumably leads to a decline in portal I/G ratio. Although in the current study we did not detect an increase in plasma glucagon in the DAPA group, the I/G ratio changed significantly. More importantly, however, when SAXA was combined with DAPA, neither plasma insulin nor plasma glucagon concentrations changed and, the I/G ratio in the peripheral circulation remained constant. Similar trends were also observed when C-pep/G ratios were calculated. Despite the absence of hormonal changes, SGLT2i-induced stimulation of EGP was not attenuated. Therefore, our data are consistent with the notion that neither a decline in plasma insulin nor an elevation in plasma glucagon can explain the DAPA-induced stimulation of EGP. These findings suggest that additional factors, including, but not limited to, changes in plasma catecholamines, central nervous system mediators, and circulating substrates/hormones, may trigger the increase in EGP (27). Whatever factors are responsible for stimulating EGP following DAPA, it must be emphasized that the quantitative impact is of a major magnitude. Whereas following PCB, EGP declined by 0.44 mg/kg/min from 2.40 to 1.96 mg/kg/min, treatment with DAPA alone or in combination with SAXA was accompanied by a lesser decline in EGP of ∼0.21 mg/kg/min (from 2.47 to 2.26 and 2.48 to 2.27 mg/kg/min, respectively) (Table 2). Thus, in response to glucosuria, which in this study amounted to ∼0.61 mg/kg/min during both therapies (Table 2), the calculated compensatory elevation in EGP offset the UGE by ∼34%, regardless of whether DAPA was used alone or in combination with SAXA.

The increase in plasma FFA concentration following DAPA is consistent with previous results and contributes to the rise in plasma ketone concentration (13–17,28). In the current study, we did not observe a significant rise in plasma ketone concentration following DAPA administration. Most likely, this can be partially explained by the modest rise in plasma FFA concentration and failure of plasma glucagon concentration and the I/G ratio to increase. Nonetheless, results from indirect calorimetry indicate that rates of carbohydrate oxidation decreased by ∼40% (1.13 vs. 0.68 mg/kg/min), while lipid oxidation increased by ∼20% (0.56 vs. 0.75 mg/kg/min) after 16 weeks of DAPA therapy. This shift in substrate oxidation did not occur with the addition of SAXA, which is consistent with changes seen in plasma FFA. These data suggest that DAPA/SAXA combination treatment reduces fatty acid availability and inhibits lipid oxidation. Moreover, by preventing the decrease in rate of carbohydrate oxidation, addition of SAXA to DAPA further contributes to lowering of plasma glucose concentration.

There are some limitations to our study. The research design did not include a group treated with SAXA alone, since the primary aim was to examine whether the addition of SAXA to DAPA would augment the decline in HbA_1c and block DAPA-induced stimulation of EGP. Furthermore, the effect of monotherapy with DPP4i on glucose homeostasis has been already well studied (24,29). The majority of patients in the current study were receiving background therapy with metformin with or without sulfonylurea. However, dosages were not changed during the study, background therapies were equally represented in all three groups, and there are no known metabolic interactions between DAPA and either metformin or sulfonylureas. Finally, this study was conducted over a relatively short period of 16 weeks, and a longer treatment period will be performed to establish sustainability of our results.

In summary, addition of SAXA to DAPA in patients with type 2 diabetes for 16 weeks leads to greater improvement in glycemic control than DAPA alone. DAPA-induced glucosuria results in stimulation of EGP that persists for up to 16 weeks and offsets by one-third the amount of glucose excreted in the urine, whether used alone or in combination with SAXA. Although the increase in EGP was unaffected by the addition of SAXA, combined DAPA/SAXA treatment was accompanied by less fatty acid oxidation and more glucose oxidation, which contributes to a greater reduction in plasma glucose and potentially could reduce the risk of ketogenesis in patients with type 2 diabetes treated with an SGLT2i.

Acknowledgments. The authors acknowledge the support of the staff and faculty of the CRC of Texas Diabetes Institute–University Health System and nurse coordinator Khanh Horst for invaluable care of the study subjects. The authors are grateful to Lorrie Albarado, administrative assistant, Division of Diabetes, for help with the manuscript preparation.

Funding and Duality of Interest. Funding for this study was provided in part by AstraZeneca and the Texas Diabetes Institute–University Health System and University of Texas Health Science Center at San Antonio and in part by National Institutes of Health grant DK-24091. C.T. has received honoraria from the speakers bureaus of Boehringer Ingelheim and AstraZeneca. R.D. has received honoraria from the speakers bureaus of AstraZeneca and Novo Nordisk; is a member of the advisory boards of AstraZeneca, Novo Nordisk, Janssen Pharmaceuticals, Boehringer Ingelheim, Eli Lilly, and Intarcia; and has received grant support from AstraZeneca, Janssen Pharmaceuticals, and Novo Nordisk. E.C. has received honoraria from the speakers bureaus of Sanofi, AstraZeneca, and Boehringer Ingelheim-Eli Lilly Alliance and grants from AstraZeneca, Janssen Pharmaceuticals, and VeroScience. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. Y.Q., C.S.-H., C.T., and E.C. executed the research procedures and contributed to sample collection, laboratory analyses, and data interpretation. J.A. provided laboratory analyses and technical assistance. E.C. and R.D. designed the study; supervised the research procedures; provided clinical management and data interpretation; and wrote, reviewed, and edited the manuscript. E.C. 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 in abstract form at the 77th Scientific Sessions of the American Diabetes Association, San Diego, CA, 9–13 June 2017.