Dapagliflozin Alters Plasma Metabolome in Patients With Type 1 Diabetes During Ketosis

Rosa Yang; Kevin Cho; Max C Petersen; Kai E Jones; Gary J Patti; Janet B McGill; Samantha A DiGruccio

doi:10.1111/dom.71390

. 2026 Sep 28;28(12):11796–11807. doi: 10.1111/dom.71390

Dapagliflozin Alters Plasma Metabolome in Patients With Type 1 Diabetes During Ketosis

Rosa Yang ¹, Kevin Cho ², Max C Petersen ¹, Kai E Jones ¹, Gary J Patti ², Janet B McGill ¹, Samantha A DiGruccio ^3,^✉

PMCID: PMC13731049 PMID: 42806245

ABSTRACT

Aims

Use of sodium‐glucose cotransporter 2 inhibitors (SGLT2i) in patients with type 1 diabetes (T1D) is limited by increased risk of ketoacidosis. We aimed to determine metabolic pathways affected by SGLT2i treatment in adults with T1D during ketosis induced by insulin withdrawal.

Materials and Methods

Plasma metabolomic and lipidomic analysis by ultra‐high‐performance liquid chromatography/mass spectrometry was performed at baseline and after supervised insulin withdrawal in 17 adults with T1D treated with usual care (UC) and (DP) dapagliflozin (10 mg daily for 14 days).

Results

Plasma glucose levels were lower, and ketone levels were higher after insulin withdrawal in DP compared to UC. Pathway analysis of differentially altered metabolites revealed that DP altered amino acids and TCA cycle metabolites during ketosis. DP treatment increased citrate and aconitate at baseline, with a trend towards increased branched chain ketoacids and acylcarnitines. Citrate, aconitate, α‐ketoglutarate and malate were increased during insulin withdrawal after UC whereas insulin withdrawal after DP resulted in unchanged α‐ketoglutarate and malate and significantly decreased fumarate and succinate. Acetylcarnitine was significantly higher during insulin withdrawal after DP compared to UC. DP effected subtle changes in amino acids and amino acid derivatives. Symmetric dimethylarginine and pseudouridine were significantly elevated at baseline and after insulin withdrawal after DP. Lipidomics revealed increased very long TG species in DP and differences in phospholipids, particularly phosphatidylcholine.

Conclusion

In adults with T1D, dapagliflozin alters TCA cycle metabolites, amino acid metabolites and certain lipids during insulin withdrawal. Future work will validate findings and further investigate the metabolic basis of SGLT2i‐mediated ketoacidosis risk.

Keywords: clinical physiology, clinical trial, dapagliflozin, type 1 diabetes

1. Introduction

Sodium‐glucose cotransporter 2 inhibitors (SGLT2i) are highly efficacious in patients with type 2 diabetes (T2D) [1, 2, 3, 4, 5], chronic kidney disease [6, 7], and heart failure [8, 9, 10, 11]; yet, their use in patients with type 1 diabetes (T1D) has been limited due to increased risk of diabetic ketoacidosis (DKA) despite demonstrated benefits for glycaemic control [12, 13, 14, 15]. The mechanisms of SGLT2i‐mediated renal and cardiac protection are not fully understood, but appear to be independent of glucose‐lowering effects [16]. One hypothesis is a metabolic shift from glucose oxidation to fatty acid and ketone oxidation benefits renal and cardiac tissue health and function [17, 18, 19]. However, this shift towards ketone production contributes to DKA risk [20], which carries significant mortality risk [21] for patients with T1D in the setting of interruption of insulin therapy or concurrent illness. Despite the development of insulin analogs and delivery devices, only an estimated 21.7% of adults with T1D achieve the recommended HbA1c of < 7.0% [22, 23]. Suboptimal glycaemic control and risk of significant cardiovascular and renal complications [24, 25, 26] reiterate the need for effective adjuvant therapies [27].

A more complete understanding of the pathophysiologic mechanisms of both SGLT2i‐mediated benefits and ketosis risk in patients with T1D will inform the safe use of SGLT2i in patients with T1D. SGLT2i may reduce inflammation [28] and impact choline [29], branched chain amino acid (BCAA) [18], and lipid metabolism, and TCA metabolites [30, 31]. Metabolomic analyses of urine and plasma from patients with T2D treated with SGLT2i revealed changes in ketone body production, BCAA and lysine catabolism, short‐chain acylcarnitines and TCA cycle metabolites [30]. Increased levels of aconitate and fumarate suggested TCA cycle activation in patients with T2D after 4 weeks of daily empagliflozin [32]; however, decreased levels of succinate, fumarate and malate were observed in patients with T2D after 12 weeks of daily dapagliflozin [30]. Increased lipid and TCA cycle metabolites including pyruvate, α‐ketoglutarate, and succinate were reported in patients with T1D under clamped hyperinsulinemic‐euglycemic conditions after 8 weeks of empagliflozin [31]. However, the metabolomic effects of SGLT2i during ketosis in adults with T1D are unknown.

We evaluated SGLT2i‐mediated changes in plasma metabolites in adults with T1D in the basal, postabsorptive state and during ketosis induced by insulin withdrawal using untargeted, relative quantification metabolomics followed by absolute quantification of certain metabolites. Evaluating the metabolomic effects of SGLT2i in patients with T1D during ketosis induced by insulin withdrawal allows us to mimic the real‐world scenario of insulin delivery disruption faced by patients with T1D. Observing metabolite changes during insulin withdrawal in participants receiving either usual care or an SGLT2i allows us to compare and contrast metabolic pathways activated by insulin withdrawal that may be further altered by SGLT2i. Understanding mechanisms of increased DKA risk observed with SGLT2i through metabolomic profiling may reveal novel pathways underlying both the therapeutic benefits and ketotic complications of SGLT2 inhibition, helping guide their safe and more personalized use as effective adjuvant therapy in patients with T1D.

2. Materials and Methods

Twenty adults provided written informed consent prior to their participation in a single‐centre (Washington University School of Medicine, St. Louis, MO) investigator‐initiated clinical trial (Clinicaltrials.gov NCT05541484), approved by the Washington University Human Research Protection Office, IRB #202206078. Results have been previously reported [33, 34]. The study was an open‐label, randomized crossover trial in which participants completed two study periods (usual care, UC and usual care plus dapagliflozin, DP) in random order consisting of a 2‐week free‐living period followed by a supervised insulin withdrawal study in the Clinical Translational Research Unit. During study period UC, participants continued their usual insulin regimen. During study period DP, participants received dapagliflozin 10 mg daily. Insulin doses were pre‐emptively reduced by 10% (basal) and 15% (prandial) during DP.

Inclusion criteria were: (i) T1D diagnosed > 1 year prior to study entry, (ii) HbA1c ≤ 10%, (iii) stable insulin delivery method (insulin pump or multiple daily injections [MDI]), (iv) age 18–75, (v) use of Dexcom G6 or G7 CGM. Exclusion criteria were: (i) history of DKA within 6 months of study entry or > 1 episode of DKA in the past 2 years, (ii) use of SGLT2i or prior intolerance of SGLT2i, (iii) eGFR < 30 mL/min/1.73 m², (iv) adherence to a very‐low‐carbohydrate (< 90 g/day) or ketogenic diet, (v) pregnancy or plan to become pregnant within the next 3 months, (vi) history of urinary tract infection within 3 months of study entry, (vii) unwilling to avoid alcohol ingestion during the study period.

For supervised insulin withdrawal, participants were instructed to fast overnight and arrived to the research unit the following morning. Participants using MDI had basal insulin transitioned to glargine U‐100, dosed each morning, prior to the day of supervised insulin withdrawal, and were instructed to hold basal insulin the morning of the supervised insulin withdrawal study. Participants using continuous subcutaneous insulin infusion were instructed to suspend insulin delivery 1 h prior to arrival to the research unit. If initial blood glucose was < 90 or > 250 mg/dL and capillary blood β‐hydroxybutyrate (βOHB) > 1.5 mmol/L, the visit was rescheduled. During the DP period, participants took the final dose of dapagliflozin on the morning of the supervised insulin withdrawal study visit. Upon admission, an intravenous catheter was placed in the antecubital fossa for serial blood sampling during supervised insulin withdrawal. Plasma was promptly separated, aliquoted, and stored at −80°C for analysis.

Samples from three participants in the original clinical study were excluded from metabolomic analysis for the following reasons: two participants had inadequate washout of basal insulin during an insulin withdrawal day, and one participant exhibited decreasing ketone levels throughout the insulin withdrawal period for unclear reasons. Paired samples from 17 participants underwent metabolomic analysis. Period order was UC followed by DP in 10 participants and DP followed by UC in 7 participants; data from both orders were pooled for this secondary analysis.

The first blood sample, taken at the time of admission on supervised insulin withdrawal day, served as the baseline (bl) sample for metabolomics analyses. Hourly samples for point‐of‐care glucose and βOHB measurement were collected until one or more stopping criteria were met. Stopping criteria included: (i) patient request for any reason, (ii) nausea and/or vomiting, (iii) 7 h elapsed, (iv) capillary βOHB ≥ 3 mmol/L, (v) plasma glucose ≥ 400 mg/dL. When stopping criteria were met, a final plasma sample was collected. Participants were then administered insulin and monitored in the CTRU for at least 1 h after insulin administration to ensure safe discharge. For metabolomic analysis, samples obtained after insulin withdrawal at the timepoint with the highest documented plasma ketone level served as the insulin withdrawal (wd) sample. This occurred with the final sample obtained prior to insulin administration for 13 of 17 UC period study visits and 15 of 17 DP period study visits.

2.1. Glucagon Analysis

Plasma glucagon was measured by ELISA (Mercodia) at the Washington University Core Laboratory for Clinical Studies. Due to variable measurement time points and length of insulin withdrawal period between each participant, we applied a repeated measures model using Proc Mixed in SAS (V9.4, Cary NC). A spatial power covariance structure was used to account for unequally spaced repeated measures. We first included a treatment group (UC or DP) by time interaction as well as main effects for treatment, time, and sequence in the model. The interaction of treatment by time was not significant, p = 0.389.

2.2. Metabolomic Analysis, Relative Quantification

Aliquots of plasma were promptly separated and stored at −80°C. For relative quantification metabolomic analysis by ultra‐high‐performance liquid chromatography/mass spectrometry (UPLC/MS), plasma samples were extracted using CAPTIVA‐EMR Lipid 96‐well plates for solid‐phase extraction. All samples were analysed in a single batch.

2.2.1. Polar Metabolites

Fifty microlitres of plasma was transferred onto a CAPTIVA‐EMR Lipid 96‐well plate (Agilent Technologies) for solid‐phase extraction followed by addition of 200 μL of acetonitrile:methanol (1:1, v/v). Samples were mixed for 1 min at 360 rpm on an orbital shaker at room temperature. Samples were incubated for 10 min at 4°C before the addition of 150 μL of acetonitrile:methanol:water (2:2:1, v/v/v) and then mixed again at 360 rpm for 10 min at room temperature and eluted into a 96‐deepwell collection plate using a positive pressure manifold. The resulting polar eluates were stored at −80°C until LC/MS analysis.

2.2.2. Lipids

Lipids retained on the SPE material were released into a second elution plate by adding 500 μL of methyl‐tert‐butyl ether:methanol (1:1, v/v) in two steps, again using the positive pressure manifold. The combined eluates were dried under a stream of nitrogen at room temperature and reconstituted with 200 μL of 2‐propanol:methanol (1:1, v/v) prior to LC/MS analysis.

2.2.3. Ketone Body Derivatization

Fifty microlitres of the polar extract was transferred into an Eppendorf tube, followed by the addition of 50 μL of derivatization solution (35 mM 3‐NPH, 25 mM EDC, 2.5% pyridine in methanol). The mixture was vortexed for 30 s and incubated for 1 h at 4°C. After incubation, samples were centrifuged at 17000× g for 10 min at 4°C, and 50 μL of the supernatant was transferred to a new Eppendorf tube. Next, 500 μL of quenching solution (0.5 mM β‐mercaptoethanol in 0.1% formic acid in water) was added, and samples were centrifuged again. The resulting supernatant was stored at −80°C until LC/MS analysis.

2.2.4. LC/MS Analysis

UPLC/MS analyses were conducted using a Thermo Scientific Vanquish Flex UHPLC system, interfaced with a Thermo Scientific Orbitrap ID‐X Mass Spectrometer. For separation of polar metabolites, a HILICON iHILIC‐(P) Classic HILIC column (100 × 2.1 mm, 5 μm) with a HILICON iHILIC‐(P) Classic guard column (20 × 2.1 mm, 5 μm) was utilised. The mobile‐phase solvents consisted of solvent A = 20 mM ammonium bicarbonate, 2.5 μM medronic acid, 0.1% ammonium hydroxide in 95:5 water:acetonitrile and solvent B = 95:5 acetonitrile:water. The column compartment temperature was maintained at 45°C, and metabolites were eluted using a linear gradient at a flow rate of 250 mL/min as follows: 0–1 min, 90% B; 12 min, 35% B; 12.5–14.5 min, 25% B; 15 min, back to 90% B. For the separation of lipids, an Acquity Premier HSS T3 column (2.1 × 100 mm, 1.8 mm) was utilised. The mobile‐phase solvents consisted of solvent A = 10 mM ammonium formate, 5 mM medronic acid in 5:3:1 water:acetonitrile:2‐propanol and solvent B = 10 mM ammonium formate in 1:9:90 water:acetonitrile:2‐propanol. The column compartment temperature was maintained at 55°C and lipids were eluted using a linear gradient at a flow rate of 400 mL/min as follows: 0 min, 15% B; 2.5 min, 50% B; 2.6 min, 57% B; 9 min, 70% B; 9.1 min, 93% B; 11 min, 96% B; 11.1 min, 100% B; 11.1–12 min, 100% B; 12.2 min, 15% B; 12.2–16 min, 15% B. Data were acquired in positive and negative ion modes. For the separation of ketone body, SCFA and BCAA, an Acquity Premier HSS T3 column (2.1 × 100 mm, 1.8 mm) was utilised. The mobile‐phase solvents consisted of solvent A = 0.05% formic acid in water B = 0.05% formic acid in acetonitrile. The column compartment temperature was maintained at 50°C and metabolites were eluted using a linear gradient at a flow rate of 300 mL/min as follows: 0–1 min, 5% B; 2 min, 20% B; 10 min, 24% B; 20 min, 75% B; 21–22 min, 90% B; 22.5–25 min, 5% B. Data were acquired in negative ion mode.

Absolute quantification of samples was performed using the MxP 500 XL kit (Biocrates). LC/MS data were processed and analysed using XCMS, CompoundDiscoverer, and Skyline.

2.3. Statistical Analysis

Power computations, using G*Power 3.1, demonstrated > 80% power at a significance threshold of α = 0.05 to detect a 40% difference between groups for the representative BCAA metabolite isobutyrylcarnitine. Published data [35] reported that the mean ± SD for plasma isobutyrylcarnitine concentration in people with T2D was 0.06 ± 0.01 μM, and that SGLT2i treatment increased the concentration of isobutyrylcarnitine by 1.39–1.52‐fold [30, 32]. Using two‐tailed t‐tests comparing the difference between two dependent means (matched pairs), these parameters indicated that the required sample size to detect a 40% increase in isobutyrylcarnitine concentration in the SGLT2i‐treated group was ≥ 5 subjects [36].

Metabolomics data were analysed by multiple paired t‐tests with adaptive linear step‐up (Benjamini, Krieger and Yekutieli) to control the false discovery rate, FDR 5%. For heatmaps, data were normalised relative to usual care baseline. Metaboanalyst 6.0 was used for pathway analysis of differentially altered metabolites.

3. Results

Baseline characteristics for 17 participants are shown in Table 1A. The mean ± SD age was 48 ± 18 years and body mass index (BMI) was 29.8 ± 5.4 kg/m². Continuous subcutaneous insulin infusion (CSII) in hybrid closed loop systems was used by 15 (88%) of the participants. At baseline, CGM time‐in‐range (70–180 mg/dL) was 61% ± 19%, with mean haemoglobin A1c of 7.0% ± 0.9% (Table 1A). Reasons for stopping insulin withdrawal were similar between usual care (UC) and dapagliflozin (DP) (Table 1B). Severe hyperglycaemia (capillary blood glucose > 400 mg/dL) occurred in two participants in UC and none in DP, whereas marked ketosis (capillary blood βOHB > 4 mmol/L) occurred in two participants in DP and none in UC (Table 1B).

TABLE 1.

Baseline participant characteristics (A) and reason for stopping insulin withdrawal (B).

A. Baseline characteristics
N = 17
Age (year)	48 ± 18
Sex (male/female)	9/8
Race (White/Black)	16/1
Weight (kg)	99.3 (74.7–104.0)
BMI (kg/m²)	29.8 ± 5.4
Diabetes duration (year)	27 ± 16
History of DKA (%)	6 (35%)
Diabetic retinopathy	4 (24%)
Diabetic peripheral neuropathy	3 (18%)
Diabetic nephropathy	1 (6%)
Hypertension	9 (53%)
Coronary artery disease	0 (0%)
Previous stroke	0 (0%)
Average sensor glucose (mg/dL)	167 ± 26
Time in range 70–180 mg/dL (%)	61 ± 19
Insulin delivery (CSII/MDI)	15/2
Total daily insulin (units/kg)	0.69 ± 0.32
Haemoglobin A1c (%)	7.0 ± 0.9
Estimated GFR (mL/min/1.73 m²)	90 (84–94)
Urine albumin/creatinine ratio (mg/g)	11 (4–22)

B. Reason for stopping insulin withdrawal
	Usual care (UC)	Dapagliflozin (DP)
Time elapsed > 8 h	8	9
Symptomatic	3	1
Blood glucose > 400 mg/dL	2	0
Blood ketones > 4 mmol/L	0	2
Participant request	4	5
N = 17 UC and 17 DP study visits

Open in a new tab

Note: Data are expressed as mean ± SD or median (IQR) as appropriate.

Abbreviations: BMI, body mass index; CSII, continuous subcutaneous insulin infusion; DKA, diabetic ketoacidosis; GFR, glomerular filtration rate; MDI, multiple daily injections.

Capillary blood glucose concentration significantly increased from baseline to the insulin withdrawal timepoint in UC but not in DP, and the glucose concentration was significantly lower after insulin withdrawal in DP versus UC (Figure 1A). Both plasma (Figure 1B) and capillary (Figure 1C) βOHB levels significantly increased during insulin withdrawal in UC and DP and were significantly higher after insulin withdrawal in DP compared to UC. There was no difference in the number of hours after insulin withdrawal when withdrawal samples were obtained (Figure 1D).

Capillary glucose, βOHB, counter‐regulatory hormones, ketone species. (A) Capillary blood glucose levels significantly increased during insulin withdrawal in usual care (UC) (black) but not in dapagliflozin (DP) (red) and were significantly lower after insulin withdrawal in DP compared to UC. *p < 0.0001 by paired t‐test. Plasma (B) and capillary (C) βOHB levels increased during insulin withdrawal in both UC and DP and were significantly higher after insulin withdrawal in DP compared to UC. *p ≤ 0.0008 by paired t‐test. (D) Time in hours after insulin withdrawal when withdrawal samples were collected was not different between UC and DP. Data is expressed as mean ± SD, each point represents one participant. (E) Glucagon levels increased during insulin withdrawal in both UC and DP. Data are expressed as mean ± sem, data were aggregated into 2 h buckets; n varies per time bucket 6–21. (F) Cortisol values (measured by absolute quantification UPLC‐MS) were elevated in DP at both baseline (bl) and withdrawal (wd) but did not significantly increase in either arm during insulin withdrawal, *p ≤ 0.03 by paired t‐test. For (A–C) and (F), each point represents a participant with lines depicting the change from baseline (bl) to withdrawal (wd) in each individual. (G) Heatmap of ketogenic pathway metabolites, each small rectangle is an individual participant. Data are normalised to UC bl. There was no significant change during insulin withdrawal for acetone in UC or DP. βOHB, acetoacetate and acetate significantly increased during insulin withdrawal in both UC and DP. βOHB and acetoacetate were higher after insulin withdrawal in DP compared to UC (p = 0.009/q = 0.1, p = 0.02/q = 0.1, respectively).

Counter‐regulatory hormones impact the development of ketosis, and SGLT2i use has been associated with elevated glucagon levels [37]. Plasma glucagon levels tended to increase during insulin withdrawal (p = 0.0003) and were significantly higher in DP compared to UC (p = 0.01) (Figure 1E); however, inter‐individual variability was substantial (see individual participant data in Figure S1). Cortisol levels were significantly higher in DP versus UC at baseline and after insulin withdrawal; however, cortisol levels did not rise significantly during insulin withdrawal in either arm (Figure 1F).

βOHB measured by relative quantification UPLC‐MS correlated well with plasma βOHB (Pearson r = 0.98 [0.96–0.99, 95% CI], r ² = 0.955, p < 0.0001) (Figure S2A). Hexose values measured by absolute quantification UPLC‐MS correlated well with capillary blood glucose values (Pearson r = 0.92 [0.88–0.95, 95% CI], r ² = 0.85, p < 0.0001) (Figure S2B) but were systematically lower than capillary glucose levels. Ketogenic pathway metabolites measured by relative quantification UPLC‐MS including βOHB, acetoacetate and acetate were significantly increased during insulin withdrawal in both UC and DP, whereas acetone was not significantly changed during insulin withdrawal (Figure 1G). βOHB and acetoacetate were higher after insulin withdrawal in DP compared to UC (Figure 1G).

Relative quantification UPLC‐MS detected 88 metabolite and 316 lipid species. Of those, 54 metabolites and 150 lipids were significantly altered (Figure 2A,B). Pathway enrichment analysis of differentially altered metabolites between UC and DP revealed changes in the TCA cycle and amino acid metabolism (Figure 2C).

Metabolites and lipids detected by relative quantification UPLC‐MS significantly altered during insulin withdrawal. Venn diagram of metabolites (A) and lipids (B) detected by relative quantification UPLC‐MS that were significantly altered during insulin withdrawal. p < 0.05, q < 0.05. Grey, usual care (UC). Red, dapagliflozin (DP). (C) Pathway analysis (Metaboanalyst 6.0) of differentially altered metabolites (n = 20) from relative quantification UPLC‐MS.

We next examined TCA cycle metabolites measured by relative quantification UPLC‐MS. Citrate and aconitate levels were significantly increased at baseline, and α‐ketoglutarate and malate trended higher at baseline in DP compared to UC (Figure 3A, Table S1). We measured plasma acylcarnitine levels to evaluate whether increased fatty acid oxidation was associated with baseline elevations in TCA metabolites. Indeed, short‐chain acylcarnitines and several medium‐ and long‐chain species demonstrated a trend towards increased levels at baseline in DP compared to UC (Figure 3B, Table S2).

Dapagliflozin alters plasma glycolysis and TCA cycle metabolites, acylcarnitines and amino acids during insulin withdrawal. Heat maps of glycolysis and TCA cycle metabolites (A) and acylcarnitines (B) measured by relative quantification UPLC‐MS. Data shown as mean, normalised to UC bl. For (B), upper panel represents acylcarnitines measured by relative quantification UPLC‐MS after eluting with lipids (see methods) with corresponding vertical legend showing change relative to UC bl. Lower panel depicts CAR 4‐OH measured by relative quantification UPLC‐MS after eluting with polar metabolites (see methods); fold change of CAR 4‐OH after insulin withdrawal was much higher than other acylcarnitines, see horizontal legend to right of heat map. (C) Heat map of amino acids measure by absolute quantification UPLC‐MS. Data are not normalised and shown as mean log2 values (μM). bl, baseline; DP, dapagliflozin; UC, usual care; wd, insulin withdrawal.

During insulin withdrawal, acylcarnitine profiling demonstrated increased lipid oxidation in both UC and DP (Figure 3B, Table S2). Free carnitine and deoxycarnitine decreased significantly during insulin withdrawal in both UC and DP, indicating depletion of the free carnitine pool as expected during a shift to fatty acid oxidation (Table S2). Acetylcarnitine (CAR 2:0) significantly increased while propionylcarnitine (CAR 3:0) significantly decreased during insulin withdrawal in both UC and DP; butyrylcarnitine (CAR 4:0) and valerylcarnitine (CAR 5:0) were not changed in either UC or DP, while 3‐hydroxybutyrylcarnitine (CAR4‐OH) was significantly increased during insulin withdrawal in both UC and DP. CAR 2:0 and CAR4‐OH were increased to a greater degree after insulin withdrawal in DP compared to UC (Figure 3B, Table S2). Changes in medium chain acylcarnitine species were heterogeneous during insulin withdrawal, with several species decreasing only in DP but not UC (CAR 8:0, CAR 13:0, CAR 7‐DC) (Figure 3B, Table S2). Most long‐chain acylcarnitines (CAR14–CAR18) significantly increased during insulin withdrawal in both arms (Table S2).

During insulin withdrawal, we observed a shift away from glycolysis that was more pronounced in DP. Pyruvate levels significantly decreased during insulin withdrawal in DP but were unchanged during insulin withdrawal in UC (Figure 3A, Table S1). Lactate levels significantly decreased during insulin withdrawal in both UC and DP but to a greater degree in DP (Figure 3A, Table S1). During insulin withdrawal, TCA metabolites citrate, aconitate, α‐ketoglutarate and malate were significantly increased in UC, while only aconitate was increased in DP (Figure 3A, Table S1). Furthermore, during insulin withdrawal in DP, succinate and fumarate were significantly decreased (Figure 3A, Table S1).

Amino acids are important fuel sources for the TCA cycle and anaplerosis. During insulin withdrawal and subsequent ketosis, amino acid breakdown contributes to ketogenic and gluconeogenic pathways. We performed absolute quantification UPLC‐MS to determine whether dapagliflozin alters amino acid metabolism during ketosis. Analysis of samples by absolute quantification UPLC‐MS detected 60 metabolites including amino acids and 397 lipid species. Of those, 40 metabolites and 179 lipids were significantly altered (Figure S3). BCAA valine, leucine and isoleucine were significantly increased during insulin withdrawal in both UC and DP (Figure 3C, Table S3). After insulin withdrawal, leucine and isoleucine were higher in DP compared to UC (Figure 3C, Table S3). Consistent with increased BCAA catabolism, BCAA‐derived ketoacids rose with insulin withdrawal. Ketoacids α‐KIV and α‐KIC, derived from valine and leucine, respectively, were increased during insulin withdrawal in both UC and DP (Table S4). α‐KMV, derived from isoleucine, was significantly increased during insulin withdrawal in UC but did not reach significance in DP (Table S4). α‐KIC and α‐KMV levels were increased at baseline in DP compared to UC. α‐KIV and α‐KIC were detected both in non‐derivatized and derivatized relative quantification analyses while α‐KMV was detected only in derivatized analysis. Results for α‐KIV and α‐KIC were slightly discordant between derivatized and non‐derivatized analyses: α‐KIV detected in non‐derivatized analysis was significantly increased during insulin wd in DP, while α‐KIV detected in derivatized analysis was not significant (Table S4). α‐KIC detected in non‐derivatized analysis was significantly increased during insulin withdrawal in DP, while α‐KIC detected in derivatized analysis was increased but q = 0.09 (Table S4). Taken together these data suggest that DP exposure increases BCAA catabolism at baseline but that BCAA catabolism during insulin withdrawal is overall similar to UC.

The ketogenic amino acid phenylalanine increased during insulin withdrawal in UC and DP (Figure 3C, Table S3). Other ketogenic amino acids threonine, tryptophan and tyrosine decreased during insulin withdrawal in UC and DP, and threonine levels were lower after insulin withdrawal in DP compared to UC (Table S3).

Glycine, glutamate, aspartate, proline, hydroxyproline, serine, citrulline, sarcosine and alanine decreased significantly during insulin withdrawal in UC and DP, while asparagine significantly decreased during insulin withdrawal in DP but not UC (Figure 3C, Table S3), possibly reflecting their diversion into gluconeogenic and other pathways.

While arginine was unchanged during insulin withdrawal in both UC and DP, several arginine metabolites were significantly changed including homoarginine and symmetric dimethylarginine (SDMA), which were both significantly increased during insulin withdrawal in both UC and DP (Table S5). SDMA was significantly increased at baseline and after insulin withdrawal in DP compared to UC (Table S5).

Additional amino acid–derived and gut microbial metabolites showed consistent changes with insulin withdrawal in both UC and DP, including decreases in methylhistidines, trans‐4‐hydroxyproline, p‐cresol sulphate and multiple indole derivatives of Trp, alongside increases in α‐aminobutyric acid (a byproduct of Met/Thr/Ser/Gly) (Table S5). α‐Aminobutyric acid trended towards increased levels in DP at baseline and after insulin withdrawal compared to UC (Table S5), suggesting that DP shifts amino acid metabolism.

Several nucleosides, purines and pyrimidines measured by relative quantification UPLC‐MS increased during insulin withdrawal in both UC and DP, including xanthine, uridine and pseudouridine (Table S6). Methylguanine was increased during insulin withdrawal in UC but not DP. Xanthine and methylguanine trended towards increased at baseline in DP, while pseudouridine was significantly increased at baseline in DP compared to UC. Pseudouridine was increased after insulin withdrawal in DP compared to UC (Table S5).

We next examined lipidomics data. Nearly all species of plasma triglycerides measured by relative quantification UPLC‐MS significantly decreased during insulin withdrawal; however, the very long chain TG 58:9 increased during insulin withdrawal in both UC and DP (Figure 4A, Table S7). DP treatment resulted in a shift towards increased levels of very long chain TG species TG 56:9 and TG 56:7 during insulin withdrawal that were unchanged during UC insulin withdrawal (Figure 4A, Table S7). Absolute quantification analysis allowed for detection of more specific triglyceride sub‐species and again showed that DP treatment resulted in increased levels of very long chain TG species during insulin withdrawal that were not changed during UC insulin withdrawal (Table S7). Most diacylglycerol species measured by relative quantification UPLC‐MS decreased during insulin withdrawal in both UC and DP; however, DG 16:0_22:5, DG 18:1_20:4 and DG 18:2_20:4 decreased only in UC (Figure 4B, Table S7). Overall, DP treatment appears to attenuate the magnitude of decrease in TG and DG species during insulin withdrawal (Figure 4A,B).

Dapagliflozin alters plasma lipidomics measured by relative quantification UPLC‐MS during insulin withdrawal. Heat map of triglycerides (A), diacylglycerides (B), ceramides (C) and phospholipids (D) measured by relative quantification UPLC‐MS that were significantly altered during insulin withdrawal. Data expressed as mean fold change from baseline to withdrawal in usual care (UC) and dapagliflozin (DP) study arms. * indicates p value < 0.05, q < 0.05; † indicates p value < 0.05, q = 0.05–0.08.

Cholesteryl esters and sphingomyelins were mostly unchanged during insulin withdrawal (Table S7). A subset of ceramide species was significantly increased during insulin withdrawal in both UC and DP (Figure 4C).

Phospholipid species measured by relative quantification UPLC‐MS showed dynamic changes during insulin withdrawal. Lysophosphatidylcholine species significantly decreased during insulin withdrawal in UC with a higher magnitude of decrease in DP, while phosphatidylcholine species significantly increased during insulin withdrawal in UC but were unchanged in DP (Figure 4D). Lysophosphatidylethanolamine and phosphatidylethanolamine species were similarly decreased while phosphatidylinositide species were similarly increased during insulin withdrawal in UC and DP (Figure 4D).

4. Discussion

Our goal was to utilise unbiased metabolomics to determine pathways that contribute to SGLT2i‐mediated ketosis and DKA in patients with T1D. Our study is the first to obtain plasma samples during ketosis due to insulin withdrawal, reflecting real‐world scenarios of insulin deficiency faced by patients due to missed insulin dosing or pump malfunction.

We observed lower glucose and higher ketone levels during insulin withdrawal in addition to increased levels at baseline of ketones, acylcarnitine species and BCAA‐derived ketoacids after 2 weeks of dapagliflozin, reflecting increased β‐oxidation and BCAA catabolism, respectively. Despite significantly increased baseline levels of citrate and aconitate in DP, metabolites distal to the α‐ketoglutarate juncture were decreased during insulin withdrawal in DP likely due to inhibition of α‐ketoglutarate dehydrogenase by a highly reduced mitochondrial state produced by unrestrained fatty acid oxidation. During insulin withdrawal, increased BCAA and decreased gluconeogenic amino acid levels suggest that SGLT2i contribute to a state of high muscle proteolysis and increased diversion of amino acids towards gluconeogenesis. During insulin withdrawal, increased levels of triglyceride species with higher total carbon number in DP may reflect SGLT2i‐mediated changes in lipoprotein lipase expression [38, 39]. Rat models of T2D demonstrated SGLT2i‐mediated alterations in hepatic [40] and myocardial [41] phospholipid profiles. Indeed, during insulin withdrawal in DP, we observed a higher magnitude of decrease in lysophosphatidylcholine species but without significant increases in phosphatidylcholine species, perhaps reflecting decreased LPCAT3 (lysophosphatidylcholine acyltransferase 3) activity. Others have observed an SGLT2i‐mediated increase in LDL cholesterol perhaps due to impacts on HMGCS1/2 [42, 43]; we did not measure serum LDL‐cholesterol in our study, but these findings should be evaluated in patients with T1D. Our metabolomic analysis recapitulates other findings that explain increased ketosis risk in patients with T1D treated with SGLT2i including increased β‐oxidation (increased acylcarnitines, increased BCAA‐derived ketoacids, increased counter‐regulatory hormones). Our data implicate other pathways in SGLT2i‐mediated ketosis in patients with T1D including lipid metabolism (alterations in high carbon TG species, lysophosphatidylcholine) and the TCA cycle.

Limitations of our study include the exploratory nature of our data collection. Sample size was relatively small (n = 17); however, study design permitted paired analysis of baseline and insulin withdrawal samples and UC and DP from the same subject. Use of plasma samples limits the ability to interpret findings at the level of specific organs and tissues. Minor differences in the statistical significance of BCAA‐derived keto‐acid responses between the underivatized and derivatized assays likely reflect workflow‐specific differences in sample preparation, derivatization efficiency, matrix effects, analytical sensitivity and measurement precision. For metabolites measured by both assays, the direction and magnitude of change were generally concordant, with discrepancies largely limited to whether individual comparisons reached statistical significance in this modestly sized cohort. Thus, the convergent findings from two orthogonal analytical approaches support the conclusion that DP exposure increases BCAA catabolism at baseline but that BCAA catabolism during insulin withdrawal is overall similar to UC. A leading hypothesis for the mechanism of SGLT2i induced ketosis is that glycosuria lowers blood glucose levels, resulting in lower insulin and higher glucagon levels, which drive lipolysis and ketogenesis [44]. Whether this mechanism is operative in the setting of T1D where counterregulatory hormones are often dysregulated is still an open question. Seven‐day exposure of patients with T1D to dapagliflozin did not alter glucagon levels [45]. Fasting glucagon levels after 4 weeks of dapagliflozin were slightly higher in patients with T1D, with a median within‐participant difference of only 2.75 pg/mL [46]. We observed similar glucagon levels at baseline between UC and DP. Glucagon levels were higher in DP compared to UC during insulin withdrawal; however, the individual results were highly variable. Cortisol levels were elevated in DP similar to other studies [46] but unchanged during insulin withdrawal in both arms. We did not measure epinephrine or norepinephrine levels but no difference was observed in these hormones in other studies [46]. These data challenge whether counter‐regulatory hormones changes are the key driver of SGLT2i‐mediated ketosis. Although we observed a high degree of inter‐individual variability, glucagon and cortisol levels were elevated in DP compared to control, indicating that these counter‐regulatory hormones could contribute to SGLT2i‐mediated ketosis during insulin withdrawal. Our data suggest that in addition to elevations in counter‐regulatory hormones, SGLT2i mediate changes in β‐oxidation pathways, other lipid metabolism pathways and the TCA cycle to increase ketosis.

TCA cycle flux regulates fuel utilisation in cells and tissues, and TCA cycle metabolites impact cellular signalling via epigenetic regulation, immune modulation and stem cell fate [47]. TCA cycle intermediates are dysregulated in T1D. Plasma fumarate levels were decreased in patients with T1D compared to healthy controls and correlated with autonomic neuropathy risk [48]. Kidney samples from T1D subjects revealed lower levels of α‐ketoglutarate, succinate and fumarate compared to controls, suggesting mitochondrial dysfunction and reduced oxidative metabolism that contribute to early diabetic kidney injury [49]. While we observed no change in succinate and fumarate at baseline, reductions in these TCA cycle intermediates during insulin withdrawal suggest a high redox state mediated by DP, which may have downstream impacts on cellular signalling.

Evaluation of a limited set of 45 plasma [31], 68 urine metabolites [31] and 32 amino acids in urine [50] in subjects with T1D under clamped euglycemia versus hyperglycaemia conditions before and after treatment with empagliflozin (25 mg daily) for 8 weeks revealed significant increases in plasma but not urine TCA cycle metabolites via dimensional reduction with principal component analysis [31]. Succinate, citrate and α‐ketoglutarate were the only TCA cycle metabolites measured. Succinate was significantly different at euglycemia after SGLT2i but was unchanged between euglycemia to hyperglycaemia in either group. Citrate was not significantly different between euglycemia and hyperglycaemia or control and empagliflozin treatment. α‐ketoglutarate was significantly different at euglycemia after SGLT2i and between euglycemia and hyperglycaemia in control but not empagliflozin group. Importantly, βOHB was not significantly different between euglycemic and hyperglycaemic clamp states in either the control or empagliflozin group, but was significantly higher in the empagliflozin group both at euglycemic and hyperglycaemic clamp states. Our data represent a state of hyperglycaemia and mild ketosis (UC wd) and normoglycemia and mild–moderate ketosis (DP wd). Urinary amino acid analysis showed that during acute hyperglycaemia, empagliflozin significantly attenuated increases in BCAA biosynthesis [50]. During insulin withdrawal, we observed increases in plasma BCAA and BCAA‐derived ketoacids, likely reflecting SGLT2i‐mediated effects in skeletal muscle as this is the primary local of BCAA catabolism. Indeed, SGLT2i were found to reduce phosphorylation‐mediated inactivation of BCKDH and impact BCAA metabolism in a mouse model of T2D [51].

Substantial metabolomic analyses of various tissues have been completed in murine models of T2D. Dapagliflozin treatment reduced accumulation of glycolytic pathway intermediates in adipose tissue of diabetic mice, but TCA cycle intermediates were unchanged [52]. Elevated acylcarnitine and βOHB levels and gene expression data suggested SGLT2i mediated enhanced mitochondrial fatty‐acid oxidation and ketone body uptake and catabolism in adipose tissue [52]. Luseogliflozin reduced mitochondrial ROS generation in islet β cells and increased TCA metabolites pyruvate, citrate, cis‐aconitate, fumarate, and malate in db/db mice [53]. Studies of SGLT2i in Akimba [54] and Akita [55] murine models of T1D revealed increased glucagon expression in islets [54] and suppressed macrophage infiltration and oxidative stress in kidney [55], respectively, with overall improved glycaemic control.

Given the efficacy in preventing T2D complications, many assume adjunct SGLT2i therapy will confer similar benefits in T1D. Long term CVD risk data in T1D on SGLT2i is lacking, but meta‐analysis of the DEPICT and inTandem trials (dapagliflozin and sotagliflozin as adjunct therapy to insulin in T1D, respectively) showed no improvement in risk ratio for cardiovascular events [56]; however the number of cardiovascular events was small and trial duration was only 1 year. SGLT2i alter levels of arginine‐related cardiovascular and renal biomarkers in patients with T2D [57] and T1D (our study) in a direction unfavourable to clinical risk. Higher SDMA levels positively correlate with diabetic retinopathy [58] and negatively correlate with eGFR [59] in patients with T1D and T2D. Murine models of CKD [60] and T2D [61, 62] demonstrate conflicting data regarding SGLT2i‐mediated alterations in SDMA; however, human data [57] consistently demonstrates increased SDMA levels after SGLT2i treatment. Pseudouridine levels in urine [63] and plasma [64] were correlated with progression of diabetic nephropathy in patients with T1D. While a murine model of T2D showed no changes in pseudouridine after empagliflozin [65], additional human data, particularly in T1D, are needed to understand the significance and pathophysiology of SGLT2i‐mediated alterations in pseudouridine. Similar to SDMA, it is unclear whether DP‐induced elevations of these biomarkers are clinically relevant or simply represent DP‐mediated changes in renal function.

SGLT2i remain an attractive class of adjunct therapy in T1D to improve glycaemic control, reduce total insulin requirements and potentially prevent complications. Understanding how to mitigate DKA risk is imperative. Continuous ketone monitoring is an exciting technology under investigation that may enable more patients to safely use SGLT2i. Besides ketones, our findings suggest that acylcarnitines, branched‐chain amino acid–derived ketoacids, and TCA cycle intermediates such as citrate or aconitate may serve as candidate biomarkers of enhanced ketogenesis and impending ketoacidosis during SGLT2 inhibitor therapy. However, further elucidation of underlying molecular mechanisms is of paramount importance. Future studies should be aimed at understanding tissue specific impacts of SGLT2i and mitochondrial function, utilising metabolic flux analysis via stable isotope tracers. Correlation of metabolite changes with clinical endpoints such as changes in GFR, albuminuria and CVD risk will help validate SGLT2i‐mediated prevention of complications of T1D.

Funding

This work was supported by Washington University in St Louis Institute of Clinical and Translational Sciences (CTRFP1712), Washington University Diabetes Research Center (National Institutes of Health P30 DK020579), Breakthrough T1D (formerly JDRF; 2‐SRA‐2022_1190‐M‐B/P22‐03211). Individuals were supported by National Institutes of Health funding as follows: S.A.D. (K08DK142012, K12DK133995), R.Y. (T32DK007120), and M.C.P. (K23 HL179482, UL1TR002345, including subaward KL2TR002346).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Plasma glucagon levels for individual participants during insulin withdrawal. Plasma glucagon levels during insulin withdrawal for each individual participant. Black line represents usual care arm, red line represents dapagliflozin arm.

Figure S2: Correlation between clinical chemistry and UPLC‐MS results for βOHB and glucose. Clinical lab values and relative quantification UPLC‐MS derived values for β‐hydroxybutyrate (A) and absolute quantification UPLC‐MS derived values for glucose (B) were well correlated. Each point represents one sample, n = 17 for each usual care baseline (black open circles) and withdrawal (black circles) and dapagliflozin baseline (red open circles) and withdrawal (black circles).

Figure S3: Metabolites and lipids detected by absolute quantification UPLC‐MS significantly altered during insulin withdrawal. Venn diagram of metabolites (A) and lipids (B) detected in absolute quantification analysis with Biocrates kit that were significantly altered during insulin withdrawal. p < 0.05, q < 0.05. Grey, usual care (UC). Red, dapagliflozin (DP).

Table S1: TCA cycle metabolites measured by relative quantification UPLC‐MS.

Table S2: Acylcarnitines measured by UPLC‐MS.

Table S3: Amino acids measured by absolute quantification UPLC‐MS.

Table S4: Ketoacids measured by untargeted, relative quantification UPLC‐MS.

Table S5: Amino Acids metabolites, gut microbe breakdown byproducts measured by absolute quantification UPLC‐MS.

Table S6: Nucleosides, purines, pyrimidines measured by untargeted, relative quantification UPLC‐MS.

Table S7: Lipidomics (excel file).

DOM-28-11796-s001.pdf^{(621.5KB, pdf)}

Acknowledgements

We thank the Washington University Core Lab for Clinical Studies for their assistance in sample analysis. We appreciate the assistance of Julia Huecker from the Center for Biostatistics and Data Science, Washington University in St. Louis, School of Medicine for statistical analysis of plasma glucagon levels.

Yang R., Cho K., Petersen M. C., et al., “Dapagliflozin Alters Plasma Metabolome in Patients With Type 1 Diabetes During Ketosis,” Diabetes, Obesity and Metabolism 28, no. 12 (2026): 11796–11807, 10.1111/dom.71390.

Handling Editor: Nigel Irwin

Data Availability Statement

The data that support the findings of this study are available in the Supporting Information of this article and are available from the corresponding author upon reasonable request.

References

1. Zinman B., Wanner C., Lachin J. M., et al., “Empagliflozin, Cardiovascular Outcomes, and Mortality in Type 2 Diabetes,” New England Journal of Medicine 373 (2015): 2117–2128. [DOI] [PubMed] [Google Scholar]
2. Rådholm K., Figtree G., Perkovic V., et al., “Canagliflozin and Heart Failure in Type 2 Diabetes Mellitus: Results From the CANVAS Program,” Circulation 138 (2018): 458–468. [DOI] [PMC free article] [PubMed] [Google Scholar]
3. Neal B., Perkovic V., Matthews D. R., et al., “Rationale, Design and Baseline Characteristics of the CANagliflozin cardioVascular Assessment Study‐Renal (CANVAS‐R): A Randomized, Placebo‐Controlled Trial,” Diabetes, Obesity & Metabolism 19 (2017): 387–393. [DOI] [PMC free article] [PubMed] [Google Scholar]
4. Zelniker T. A., Wiviott S. D., Raz I., et al., “SGLT2 Inhibitors for Primary and Secondary Prevention of Cardiovascular and Renal Outcomes in Type 2 Diabetes: A Systematic Review and Meta‐Analysis of Cardiovascular Outcome Trials,” Lancet (London, England) 393 (2019): 31–39. [DOI] [PubMed] [Google Scholar]
5. Wiviott S. D., Raz I., Bonaca M. P., et al., “Dapagliflozin and Cardiovascular Outcomes in Type 2 Diabetes,” New England Journal of Medicine 380 (2019): 347–357. [DOI] [PubMed] [Google Scholar]
6. Perkovic V., Jardine M. J., Neal B., et al., “Canagliflozin and Renal Outcomes in Type 2 Diabetes and Nephropathy,” New England Journal of Medicine 380 (2019): 2295–2306. [DOI] [PubMed] [Google Scholar]
7. Heerspink H. J. L., Stefánsson B. V., Correa‐Rotter R., et al., “Dapagliflozin in Patients With Chronic Kidney Disease,” New England Journal of Medicine 383 (2020): 1436–1446. [DOI] [PubMed] [Google Scholar]
8. McMurray J. J. V., Solomon S. D., Inzucchi S. E., et al., “Dapagliflozin in Patients With Heart Failure and Reduced Ejection Fraction,” New England Journal of Medicine 381 (2019): 1995–2008. [DOI] [PubMed] [Google Scholar]
9. Packer M., Anker S. D., Butler J., et al., “Cardiovascular and Renal Outcomes With Empagliflozin in Heart Failure,” New England Journal of Medicine 383 (2020): 1413–1424. [DOI] [PubMed] [Google Scholar]
10. Anker S. D., Butler J., Filippatos G., et al., “Empagliflozin in Heart Failure With a Preserved Ejection Fraction,” New England Journal of Medicine 385 (2021): 1451–1461. [DOI] [PubMed] [Google Scholar]
11. Voors A. A., Angermann C. E., Teerlink J. R., et al., “The SGLT2 Inhibitor Empagliflozin in Patients Hospitalized for Acute Heart Failure: A Multinational Randomized Trial,” Nature Medicine 28 (2022): 568–574. [DOI] [PMC free article] [PubMed] [Google Scholar]
12. Rodbard H. W., Peters A. L., Slee A., Cao A., Traina S. B., and Alba M., “The Effect of Canagliflozin, a Sodium Glucose Cotransporter 2 Inhibitor, on Glycemic End Points Assessed by Continuous Glucose Monitoring and Patient‐Reported Outcomes Among People With Type 1 Diabetes,” Diabetes Care 40 (2017): 171–180. [DOI] [PubMed] [Google Scholar]
13. Garcia‐Tirado J., Farhy L., Nass R., et al., “Automated Insulin Delivery With SGLT2i Combination Therapy in Type 1 Diabetes,” Diabetes Technology & Therapeutics 24 (2022): 461–470. [DOI] [PMC free article] [PubMed] [Google Scholar]
14. Wolfsdorf J. I. and Ratner R. E., “SGLT Inhibitors for Type 1 Diabetes: Proceed With Extreme Caution,” Diabetes Care 42 (2019): 991–993. [DOI] [PubMed] [Google Scholar]
15. Buse J. B., Garg S. K., Rosenstock J., et al., “Sotagliflozin in Combination With Optimized Insulin Therapy in Adults With Type 1 Diabetes: The North American inTandem1 Study,” Diabetes Care 41 (2018): 1970–1980. [DOI] [PMC free article] [PubMed] [Google Scholar]
16. Schnell O., Valensi P., Standl E., and Ceriello A., “Comparison of Mechanisms and Transferability of Outcomes of SGLT2 Inhibition Between Type 1 and Type 2 Diabetes,” Endocrinology, Diabetes & Metabolism 3 (2020): e00129. [DOI] [PMC free article] [PubMed] [Google Scholar]
17. Xi Y., Chen D., Dong Z., et al., “Multi‐Omics Insights Into Potential Mechanism of SGLT2 Inhibitors Cardiovascular Benefit in Diabetic Cardiomyopathy,” Frontiers in Cardiovascular Medicine 9 (2022): 999254. [DOI] [PMC free article] [PubMed] [Google Scholar]
18. Selvaraj S., Fu Z., Jones P., et al., “Metabolomic Profiling of the Effects of Dapagliflozin in Heart Failure With Reduced Ejection Fraction: DEFINE‐HF,” Circulation 146 (2022): 808–818. [DOI] [PMC free article] [PubMed] [Google Scholar]
19. Gao Y.‐M., Feng S. T., Wen Y., Tang T. T., Wang B., and Liu B. C., “Cardiorenal Protection of SGLT2 Inhibitors‐Perspectives From Metabolic Reprogramming,” eBioMedicine 83 (2022): 104215. [DOI] [PMC free article] [PubMed] [Google Scholar]
20. Patoulias D., Manafis A., Mitas C., et al., “Sodium‐Glucose Cotransporter 2 Inhibitors and the Risk of Diabetic Ketoacidosis; From Pathophysiology to Clinical Practice,” Cardiovascular & Hematological Disorders Drug Targets 18 (2018): 139–146. [DOI] [PubMed] [Google Scholar]
21. Ramphul K. and Joynauth J., “An Update on the Incidence and Burden of Diabetic Ketoacidosis in the U.S,” Diabetes Care 43 (2020): e196–e197. [DOI] [PubMed] [Google Scholar]
22. Foster N. C., Beck R. W., Miller K. M., et al., “State of Type 1 Diabetes Management and Outcomes From the T1D Exchange in 2016‐2018,” Diabetes Technology & Therapeutics 21 (2019): 66–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
23. Redondo M. J., Libman I., Maahs D. M., et al., “The Evolution of Hemoglobin A1c Targets for Youth With Type 1 Diabetes: Rationale and Supporting Evidence,” Diabetes Care 44 (2021): 301–312. [DOI] [PMC free article] [PubMed] [Google Scholar]
24. Maahs D. M., Daniels S. R., De Ferranti S. D., et al., “Cardiovascular Disease Risk Factors in Youth With Diabetes Mellitus: A Scientific Statement From the American Heart Association,” Circulation 130 (2014): 1532–1558. [DOI] [PubMed] [Google Scholar]
25. Pambianco G., Costacou T., Ellis D., Becker D. J., Klein R., and Orchard T. J., “The 30‐Year Natural History of Type 1 Diabetes Complications: The Pittsburgh Epidemiology of Diabetes Complications Study Experience,” Diabetes 55 (2006): 1463–1469. [DOI] [PubMed] [Google Scholar]
26. Costacou T. and Orchard T. J., “Cumulative Kidney Complication Risk by 50 Years of Type 1 Diabetes: The Effects of Sex, Age, and Calendar Year at Onset,” Diabetes Care 41 (2018): 426–433. [DOI] [PMC free article] [PubMed] [Google Scholar]
27. Cai X., Lin C., Yang W., Nie L., and Ji L., “Non‐Insulin Antidiabetes Treatment in Type 1 Diabetes Mellitus: A Systematic Review and Meta‐Analysis,” Diabetes and Metabolism Journal 45 (2021): 312–325. [DOI] [PMC free article] [PubMed] [Google Scholar]
28. Chen H., Birnbaum Y., Ye R., Yang H. C., Bajaj M., and Ye Y., “SGLT2 Inhibition by Dapagliflozin Attenuates Diabetic Ketoacidosis in Mice With Type‐1 Diabetes,” Cardiovascular Drugs and Therapy 36 (2022): 1091–1108. [DOI] [PubMed] [Google Scholar]
29. Xu M., Zheng J., Hou T., et al., “SGLT2 Inhibition, Choline Metabolites, and Cardiometabolic Diseases: A Mediation Mendelian Randomization Study,” Diabetes Care 45 (2022): 2718–2728. [DOI] [PMC free article] [PubMed] [Google Scholar]
30. Mulder S., Hammarstedt A., Nagaraj S. B., et al., “A Metabolomics‐Based Molecular Pathway Analysis of How the Sodium‐Glucose Co‐Transporter‐2 Inhibitor Dapagliflozin May Slow Kidney Function Decline in Patients With Diabetes,” Diabetes, Obesity & Metabolism 22 (2020): 1157–1166. [DOI] [PMC free article] [PubMed] [Google Scholar]
31. Liu H., Sridhar V. S., Montemayor D., et al., “Changes in Plasma and Urine Metabolites Associated With Empagliflozin in Patients With Type 1 Diabetes,” Diabetes, Obesity & Metabolism 23 (2021): 2466–2475. [DOI] [PubMed] [Google Scholar]
32. Kappel B. A., Lehrke M., Schütt K., et al., “Effect of Empagliflozin on the Metabolic Signature of Patients With Type 2 Diabetes Mellitus and Cardiovascular Disease,” Circulation 136 (2017): 969–972. [DOI] [PubMed] [Google Scholar]
33. Jones K. E., Petersen M. C., Markov A. M., et al., “Breath Acetone Correlates With Capillary β‐Hydroxybutyrate in Type 1 Diabetes,” Journal of Diabetes Science and Technology 20 (2025): 19322968251334640, 10.1177/19322968251334640. [DOI] [PMC free article] [PubMed] [Google Scholar]
34. Petersen M. C., Jones K. E., Markov A. M., et al., “Effect of Dapagliflozin on Blood and Breath Ketones During Supervised Insulin Withdrawal in Adults With Type 1 Diabetes: A Randomized Crossover Trial,” Diabetes, Obesity & Metabolism 27 (2025): 3124–3131. [DOI] [PMC free article] [PubMed] [Google Scholar]
35. Adams S. H., Hoppel C. L., Lok K. H., et al., “Plasma Acylcarnitine Profiles Suggest Incomplete Long‐Chain Fatty Acid Beta‐Oxidation and Altered Tricarboxylic Acid Cycle Activity in Type 2 Diabetic African‐American Women,” Journal of Nutrition 139 (2009): 1073–1081. [DOI] [PMC free article] [PubMed] [Google Scholar]
36. Stancliffe E., Schwaiger‐Haber M., Sindelar M., Murphy M. J., Soerensen M., and Patti G. J., “An Untargeted Metabolomics Workflow That Scales to Thousands of Samples for Population‐Based Studies,” Analytical Chemistry 94 (2022): 17370–17378. [DOI] [PMC free article] [PubMed] [Google Scholar]
37. Bonner C., Kerr‐Conte J., Gmyr V., et al., “Inhibition of the Glucose Transporter SGLT2 With Dapagliflozin in Pancreatic Alpha Cells Triggers Glucagon Secretion,” Nature Medicine 21 (2015): 512–517. [DOI] [PubMed] [Google Scholar]
38. Basu D., Huggins L. A., Scerbo D., et al., “Mechanism of Increased LDL (Low‐Density Lipoprotein) and Decreased Triglycerides With SGLT2 (Sodium‐Glucose Cotransporter 2) Inhibition,” Arteriosclerosis, Thrombosis, and Vascular Biology 38 (2018): 2207–2216. [DOI] [PMC free article] [PubMed] [Google Scholar]
39. Kralova E., Marusakova M., Hadova K., Krenek P., and Klimas J., “Dapagliflozin Elevates Plasma High‐Density Lipoprotein Levels and Influences Visceral Fat Gene Expression in Streptozotocin‐Induced Diabetes Mellitus,” Journal of Pharmacy and Pharmacology 73 (2021): 778–784. [DOI] [PubMed] [Google Scholar]
40. Aragón‐Herrera A., Otero‐Santiago M., Anido‐Varela L., et al., “The Treatment With the SGLT2 Inhibitor Empagliflozin Modifies the Hepatic Metabolome of Male Zucker Diabetic Fatty Rats Towards a Protective Profile,” Frontiers in Pharmacology 13 (2022): 827033. [DOI] [PMC free article] [PubMed] [Google Scholar]
41. Miklankova D., Markova I., Hüttl M., and Malinska H., “Empagliflozin Alters Lipid Metabolism in the Myocardium and Liver in a Prediabetes Model With Severe Dyslipidemia,” Frontiers in Pharmacology 15 (2024): 1393946. [DOI] [PMC free article] [PubMed] [Google Scholar]
42. Briand F., Mayoux E., Brousseau E., et al., “Empagliflozin, via Switching Metabolism Toward Lipid Utilization, Moderately Increases LDL Cholesterol Levels Through Reduced LDL Catabolism,” Diabetes 65 (2016): 2032–2038. [DOI] [PubMed] [Google Scholar]
43. Bechmann L. E., Emanuelsson F., Nordestgaard B. G., and Benn M., “SGLT2‐Inhibition Increases Total, LDL, and HDL Cholesterol and Lowers Triglycerides: Meta‐Analyses of 60 Randomized Trials, Overall and by Dose, Ethnicity, and Drug Type,” Atherosclerosis 394 (2024): 117236. [DOI] [PubMed] [Google Scholar]
44. Ferrannini E., Baldi S., Frascerra S., et al., “Shift to Fatty Substrate Utilization in Response to Sodium‐Glucose Cotransporter 2 Inhibition in Subjects Without Diabetes and Patients With Type 2 Diabetes,” Diabetes 65 (2016): 1190–1195. [DOI] [PubMed] [Google Scholar]
45. Gübeli A., Steiner N., Limacher A., Mathis D., Melmer A., and Laimer M., “Dapagliflozin's Impact on Hormonal Regulation and Ketogenesis in Type 1 Diabetes: A Randomised Controlled Crossover Trial,” Diabetologia 68 (2025): 2116–2125. [DOI] [PMC free article] [PubMed] [Google Scholar]
46. Boeder S. C., Gregory J. M., Giovannetti E. R., and Pettus J. H., “SGLT2 Inhibition Increases Fasting Glucagon but Does Not Restore the Counterregulatory Hormone Response to Hypoglycemia in Participants With Type 1 Diabetes,” Diabetes 71 (2022): 511–519. [DOI] [PMC free article] [PubMed] [Google Scholar]
47. Martínez‐Reyes I. and Chandel N. S., “Mitochondrial TCA Cycle Metabolites Control Physiology and Disease,” Nature Communications 11 (2020): 102. [DOI] [PMC free article] [PubMed] [Google Scholar]
48. Mathew A. V., Jaiswal M., Ang L., Michailidis G., Pennathur S., and Pop‐Busui R., “Impaired Amino Acid and TCA Metabolism and Cardiovascular Autonomic Neuropathy Progression in Type 1 Diabetes,” Diabetes 68 (2019): 2035–2044. [DOI] [PMC free article] [PubMed] [Google Scholar]
49. Choi Y. J., Richard G., Zhang G., et al., “Attenuated Kidney Oxidative Metabolism in Young Adults With Type 1 Diabetes,” Journal of Clinical Investigation 134 (2024): e183984. [DOI] [PMC free article] [PubMed] [Google Scholar]
50. Kugathasan L., Sridhar V. S., Lovblom L. E., et al., “Interactive Effects of Empagliflozin and Hyperglycemia on Urinary Amino Acids in Individuals With Type 1 Diabetes,” Diabetes 73 (2024): 401–411. [DOI] [PMC free article] [PubMed] [Google Scholar]
51. Sakamoto M., Hasuzawa N., Wang L., et al., “Sodium‐Glucose Cotransporter 2 Inhibitors—But Not Insulin—Enhance Renal Branched‐Chain Amino Acid Catabolism,” Frontiers in Endocrinology 16 (2025): 1706838. [DOI] [PMC free article] [PubMed] [Google Scholar]
52. Nishitani S., Fukuhara A., Shin J., Okuno Y., Otsuki M., and Shimomura I., “Metabolomic and Microarray Analyses of Adipose Tissue of Dapagliflozin‐Treated Mice, and Effects of 3‐Hydroxybutyrate on Induction of Adiponectin in Adipocytes,” Scientific Reports 8 (2018): 8805. [DOI] [PMC free article] [PubMed] [Google Scholar]
53. Yamauchi Y., Nakamura A., Yokota T., et al., “Luseogliflozin Preserves the Pancreatic Beta‐Cell Mass and Function in db/db Mice by Improving Mitochondrial Function,” Scientific Reports 12 (2022): 9740. [DOI] [PMC free article] [PubMed] [Google Scholar]
54. Herat L. Y., Matthews J. R., Hibbs M., Rakoczy E. P., Schlaich M. P., and Matthews V. B., “SGLT1/2 Inhibition Improves Glycemic Control and Multi‐Organ Protection in Type 1 Diabetes,” iScience 26 (2023): 107260. [DOI] [PMC free article] [PubMed] [Google Scholar]
55. Hatanaka T., Ogawa D., Tachibana H., et al., “Inhibition of SGLT2 Alleviates Diabetic Nephropathy by Suppressing High Glucose‐Induced Oxidative Stress in Type 1 Diabetic Mice,” Pharmacology Research & Perspectives 4 (2016): e00239. [DOI] [PMC free article] [PubMed] [Google Scholar]
56. Wang S., Zhang Y., Balati A., Li B., Dai L., and Yu D., “Effect of Sodium‐Glucose Transporter 2 Inhibitors on Cardiovascular Outcomes in Type 1 Diabetes Mellitus: A Systematic Review and Meta‐Analysis of Randomised Controlled Trials,” Journal of Clinical Endocrinology and Metabolism 110, no. 7 (2025): 2071–2081, 10.1210/clinem/dgaf016. [DOI] [PubMed] [Google Scholar]
57. Gessner A., Gemeinhardt A., Bosch A., et al., “Effects of Treatment With SGLT‐2 Inhibitors on Arginine‐Related Cardiovascular and Renal Biomarkers,” Cardiovascular Diabetology 21 (2022): 4. [DOI] [PMC free article] [PubMed] [Google Scholar]
58. Abhary S., Kasmeridis N., Burdon K. P., et al., “Diabetic Retinopathy Is Associated With Elevated Serum Asymmetric and Symmetric Dimethylarginines,” Diabetes Care 32 (2009): 2084–2086. [DOI] [PMC free article] [PubMed] [Google Scholar]
59. Omran D. R. A. R., Mohamed N. A. K., Ramzy T. R. A., and Ramiz S. A., “The Role of Serum Symmetric Dimethylarginine (SDMA) as Endogenous Biomarker of Renal Impairment in Correlation With Severity and Different Stages of Diabetic Kidney Disease,” Egyptian Journal of Internal Medicine 37 (2025): 165. [Google Scholar]
60. Mishima E., Fukuda S., Kanemitsu Y., et al., “Canagliflozin Reduces Plasma Uremic Toxins and Alters the Intestinal Microbiota Composition in a Chronic Kidney Disease Mouse Model,” American Journal of Physiology. Renal Physiology 315 (2018): F824–F833. [DOI] [PubMed] [Google Scholar]
61. Adingupu D. D., Göpel S. O., Grönros J., et al., “SGLT2 Inhibition With Empagliflozin Improves Coronary Microvascular Function and Cardiac Contractility in Prediabetic Ob/Ob−/− Mice,” Cardiovascular Diabetology 18 (2019): 16. [DOI] [PMC free article] [PubMed] [Google Scholar]
62. Herat L. Y., Magno A. L., Rudnicka C., et al., “SGLT2 Inhibitor–Induced Sympathoinhibition: A Novel Mechanism for Cardiorenal Protection,” JACC: Basic to Translational Science 5 (2020): 169–179. [DOI] [PMC free article] [PubMed] [Google Scholar]
63. Mutter S., Valo E., Aittomäki V., et al., “Urinary Metabolite Profiling and Risk of Progression of Diabetic Nephropathy in 2670 Individuals With Type 1 Diabetes,” Diabetologia 65 (2022): 140–149. [DOI] [PMC free article] [PubMed] [Google Scholar]
64. Niewczas M. A., Mathew A. V., Croall S., et al., “Circulating Modified Metabolites and a Risk of ESRD in Patients With Type 1 Diabetes and Chronic Kidney Disease,” Diabetes Care 40 (2017): 383–390. [DOI] [PMC free article] [PubMed] [Google Scholar]
65. Lu Y.‐P., Zhang Z. Y., Wu H. W., et al., “SGLT2 Inhibitors Improve Kidney Function and Morphology by Regulating Renal Metabolic Reprogramming in Mice With Diabetic Kidney Disease,” Journal of Translational Medicine 20 (2022): 420. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Table S1: TCA cycle metabolites measured by relative quantification UPLC‐MS.

Table S2: Acylcarnitines measured by UPLC‐MS.

Table S3: Amino acids measured by absolute quantification UPLC‐MS.

Table S4: Ketoacids measured by untargeted, relative quantification UPLC‐MS.

Table S5: Amino Acids metabolites, gut microbe breakdown byproducts measured by absolute quantification UPLC‐MS.

Table S6: Nucleosides, purines, pyrimidines measured by untargeted, relative quantification UPLC‐MS.

Table S7: Lipidomics (excel file).

DOM-28-11796-s001.pdf^{(621.5KB, pdf)}

Data Availability Statement

The data that support the findings of this study are available in the Supporting Information of this article and are available from the corresponding author upon reasonable request.

[dom71390-bib-0001] 1. Zinman B., Wanner C., Lachin J. M., et al., “Empagliflozin, Cardiovascular Outcomes, and Mortality in Type 2 Diabetes,” New England Journal of Medicine 373 (2015): 2117–2128. [DOI] [PubMed] [Google Scholar]

[dom71390-bib-0002] 2. Rådholm K., Figtree G., Perkovic V., et al., “Canagliflozin and Heart Failure in Type 2 Diabetes Mellitus: Results From the CANVAS Program,” Circulation 138 (2018): 458–468. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0003] 3. Neal B., Perkovic V., Matthews D. R., et al., “Rationale, Design and Baseline Characteristics of the CANagliflozin cardioVascular Assessment Study‐Renal (CANVAS‐R): A Randomized, Placebo‐Controlled Trial,” Diabetes, Obesity & Metabolism 19 (2017): 387–393. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0004] 4. Zelniker T. A., Wiviott S. D., Raz I., et al., “SGLT2 Inhibitors for Primary and Secondary Prevention of Cardiovascular and Renal Outcomes in Type 2 Diabetes: A Systematic Review and Meta‐Analysis of Cardiovascular Outcome Trials,” Lancet (London, England) 393 (2019): 31–39. [DOI] [PubMed] [Google Scholar]

[dom71390-bib-0005] 5. Wiviott S. D., Raz I., Bonaca M. P., et al., “Dapagliflozin and Cardiovascular Outcomes in Type 2 Diabetes,” New England Journal of Medicine 380 (2019): 347–357. [DOI] [PubMed] [Google Scholar]

[dom71390-bib-0006] 6. Perkovic V., Jardine M. J., Neal B., et al., “Canagliflozin and Renal Outcomes in Type 2 Diabetes and Nephropathy,” New England Journal of Medicine 380 (2019): 2295–2306. [DOI] [PubMed] [Google Scholar]

[dom71390-bib-0007] 7. Heerspink H. J. L., Stefánsson B. V., Correa‐Rotter R., et al., “Dapagliflozin in Patients With Chronic Kidney Disease,” New England Journal of Medicine 383 (2020): 1436–1446. [DOI] [PubMed] [Google Scholar]

[dom71390-bib-0008] 8. McMurray J. J. V., Solomon S. D., Inzucchi S. E., et al., “Dapagliflozin in Patients With Heart Failure and Reduced Ejection Fraction,” New England Journal of Medicine 381 (2019): 1995–2008. [DOI] [PubMed] [Google Scholar]

[dom71390-bib-0009] 9. Packer M., Anker S. D., Butler J., et al., “Cardiovascular and Renal Outcomes With Empagliflozin in Heart Failure,” New England Journal of Medicine 383 (2020): 1413–1424. [DOI] [PubMed] [Google Scholar]

[dom71390-bib-0010] 10. Anker S. D., Butler J., Filippatos G., et al., “Empagliflozin in Heart Failure With a Preserved Ejection Fraction,” New England Journal of Medicine 385 (2021): 1451–1461. [DOI] [PubMed] [Google Scholar]

[dom71390-bib-0011] 11. Voors A. A., Angermann C. E., Teerlink J. R., et al., “The SGLT2 Inhibitor Empagliflozin in Patients Hospitalized for Acute Heart Failure: A Multinational Randomized Trial,” Nature Medicine 28 (2022): 568–574. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0012] 12. Rodbard H. W., Peters A. L., Slee A., Cao A., Traina S. B., and Alba M., “The Effect of Canagliflozin, a Sodium Glucose Cotransporter 2 Inhibitor, on Glycemic End Points Assessed by Continuous Glucose Monitoring and Patient‐Reported Outcomes Among People With Type 1 Diabetes,” Diabetes Care 40 (2017): 171–180. [DOI] [PubMed] [Google Scholar]

[dom71390-bib-0013] 13. Garcia‐Tirado J., Farhy L., Nass R., et al., “Automated Insulin Delivery With SGLT2i Combination Therapy in Type 1 Diabetes,” Diabetes Technology & Therapeutics 24 (2022): 461–470. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0014] 14. Wolfsdorf J. I. and Ratner R. E., “SGLT Inhibitors for Type 1 Diabetes: Proceed With Extreme Caution,” Diabetes Care 42 (2019): 991–993. [DOI] [PubMed] [Google Scholar]

[dom71390-bib-0015] 15. Buse J. B., Garg S. K., Rosenstock J., et al., “Sotagliflozin in Combination With Optimized Insulin Therapy in Adults With Type 1 Diabetes: The North American inTandem1 Study,” Diabetes Care 41 (2018): 1970–1980. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0016] 16. Schnell O., Valensi P., Standl E., and Ceriello A., “Comparison of Mechanisms and Transferability of Outcomes of SGLT2 Inhibition Between Type 1 and Type 2 Diabetes,” Endocrinology, Diabetes & Metabolism 3 (2020): e00129. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0017] 17. Xi Y., Chen D., Dong Z., et al., “Multi‐Omics Insights Into Potential Mechanism of SGLT2 Inhibitors Cardiovascular Benefit in Diabetic Cardiomyopathy,” Frontiers in Cardiovascular Medicine 9 (2022): 999254. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0018] 18. Selvaraj S., Fu Z., Jones P., et al., “Metabolomic Profiling of the Effects of Dapagliflozin in Heart Failure With Reduced Ejection Fraction: DEFINE‐HF,” Circulation 146 (2022): 808–818. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0019] 19. Gao Y.‐M., Feng S. T., Wen Y., Tang T. T., Wang B., and Liu B. C., “Cardiorenal Protection of SGLT2 Inhibitors‐Perspectives From Metabolic Reprogramming,” eBioMedicine 83 (2022): 104215. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0020] 20. Patoulias D., Manafis A., Mitas C., et al., “Sodium‐Glucose Cotransporter 2 Inhibitors and the Risk of Diabetic Ketoacidosis; From Pathophysiology to Clinical Practice,” Cardiovascular & Hematological Disorders Drug Targets 18 (2018): 139–146. [DOI] [PubMed] [Google Scholar]

[dom71390-bib-0021] 21. Ramphul K. and Joynauth J., “An Update on the Incidence and Burden of Diabetic Ketoacidosis in the U.S,” Diabetes Care 43 (2020): e196–e197. [DOI] [PubMed] [Google Scholar]

[dom71390-bib-0022] 22. Foster N. C., Beck R. W., Miller K. M., et al., “State of Type 1 Diabetes Management and Outcomes From the T1D Exchange in 2016‐2018,” Diabetes Technology & Therapeutics 21 (2019): 66–72. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0023] 23. Redondo M. J., Libman I., Maahs D. M., et al., “The Evolution of Hemoglobin A1c Targets for Youth With Type 1 Diabetes: Rationale and Supporting Evidence,” Diabetes Care 44 (2021): 301–312. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0024] 24. Maahs D. M., Daniels S. R., De Ferranti S. D., et al., “Cardiovascular Disease Risk Factors in Youth With Diabetes Mellitus: A Scientific Statement From the American Heart Association,” Circulation 130 (2014): 1532–1558. [DOI] [PubMed] [Google Scholar]

[dom71390-bib-0025] 25. Pambianco G., Costacou T., Ellis D., Becker D. J., Klein R., and Orchard T. J., “The 30‐Year Natural History of Type 1 Diabetes Complications: The Pittsburgh Epidemiology of Diabetes Complications Study Experience,” Diabetes 55 (2006): 1463–1469. [DOI] [PubMed] [Google Scholar]

[dom71390-bib-0026] 26. Costacou T. and Orchard T. J., “Cumulative Kidney Complication Risk by 50 Years of Type 1 Diabetes: The Effects of Sex, Age, and Calendar Year at Onset,” Diabetes Care 41 (2018): 426–433. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0027] 27. Cai X., Lin C., Yang W., Nie L., and Ji L., “Non‐Insulin Antidiabetes Treatment in Type 1 Diabetes Mellitus: A Systematic Review and Meta‐Analysis,” Diabetes and Metabolism Journal 45 (2021): 312–325. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0028] 28. Chen H., Birnbaum Y., Ye R., Yang H. C., Bajaj M., and Ye Y., “SGLT2 Inhibition by Dapagliflozin Attenuates Diabetic Ketoacidosis in Mice With Type‐1 Diabetes,” Cardiovascular Drugs and Therapy 36 (2022): 1091–1108. [DOI] [PubMed] [Google Scholar]

[dom71390-bib-0029] 29. Xu M., Zheng J., Hou T., et al., “SGLT2 Inhibition, Choline Metabolites, and Cardiometabolic Diseases: A Mediation Mendelian Randomization Study,” Diabetes Care 45 (2022): 2718–2728. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0030] 30. Mulder S., Hammarstedt A., Nagaraj S. B., et al., “A Metabolomics‐Based Molecular Pathway Analysis of How the Sodium‐Glucose Co‐Transporter‐2 Inhibitor Dapagliflozin May Slow Kidney Function Decline in Patients With Diabetes,” Diabetes, Obesity & Metabolism 22 (2020): 1157–1166. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0031] 31. Liu H., Sridhar V. S., Montemayor D., et al., “Changes in Plasma and Urine Metabolites Associated With Empagliflozin in Patients With Type 1 Diabetes,” Diabetes, Obesity & Metabolism 23 (2021): 2466–2475. [DOI] [PubMed] [Google Scholar]

[dom71390-bib-0032] 32. Kappel B. A., Lehrke M., Schütt K., et al., “Effect of Empagliflozin on the Metabolic Signature of Patients With Type 2 Diabetes Mellitus and Cardiovascular Disease,” Circulation 136 (2017): 969–972. [DOI] [PubMed] [Google Scholar]

[dom71390-bib-0033] 33. Jones K. E., Petersen M. C., Markov A. M., et al., “Breath Acetone Correlates With Capillary β‐Hydroxybutyrate in Type 1 Diabetes,” Journal of Diabetes Science and Technology 20 (2025): 19322968251334640, 10.1177/19322968251334640. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0034] 34. Petersen M. C., Jones K. E., Markov A. M., et al., “Effect of Dapagliflozin on Blood and Breath Ketones During Supervised Insulin Withdrawal in Adults With Type 1 Diabetes: A Randomized Crossover Trial,” Diabetes, Obesity & Metabolism 27 (2025): 3124–3131. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0035] 35. Adams S. H., Hoppel C. L., Lok K. H., et al., “Plasma Acylcarnitine Profiles Suggest Incomplete Long‐Chain Fatty Acid Beta‐Oxidation and Altered Tricarboxylic Acid Cycle Activity in Type 2 Diabetic African‐American Women,” Journal of Nutrition 139 (2009): 1073–1081. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0036] 36. Stancliffe E., Schwaiger‐Haber M., Sindelar M., Murphy M. J., Soerensen M., and Patti G. J., “An Untargeted Metabolomics Workflow That Scales to Thousands of Samples for Population‐Based Studies,” Analytical Chemistry 94 (2022): 17370–17378. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0037] 37. Bonner C., Kerr‐Conte J., Gmyr V., et al., “Inhibition of the Glucose Transporter SGLT2 With Dapagliflozin in Pancreatic Alpha Cells Triggers Glucagon Secretion,” Nature Medicine 21 (2015): 512–517. [DOI] [PubMed] [Google Scholar]

[dom71390-bib-0038] 38. Basu D., Huggins L. A., Scerbo D., et al., “Mechanism of Increased LDL (Low‐Density Lipoprotein) and Decreased Triglycerides With SGLT2 (Sodium‐Glucose Cotransporter 2) Inhibition,” Arteriosclerosis, Thrombosis, and Vascular Biology 38 (2018): 2207–2216. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0039] 39. Kralova E., Marusakova M., Hadova K., Krenek P., and Klimas J., “Dapagliflozin Elevates Plasma High‐Density Lipoprotein Levels and Influences Visceral Fat Gene Expression in Streptozotocin‐Induced Diabetes Mellitus,” Journal of Pharmacy and Pharmacology 73 (2021): 778–784. [DOI] [PubMed] [Google Scholar]

[dom71390-bib-0040] 40. Aragón‐Herrera A., Otero‐Santiago M., Anido‐Varela L., et al., “The Treatment With the SGLT2 Inhibitor Empagliflozin Modifies the Hepatic Metabolome of Male Zucker Diabetic Fatty Rats Towards a Protective Profile,” Frontiers in Pharmacology 13 (2022): 827033. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0041] 41. Miklankova D., Markova I., Hüttl M., and Malinska H., “Empagliflozin Alters Lipid Metabolism in the Myocardium and Liver in a Prediabetes Model With Severe Dyslipidemia,” Frontiers in Pharmacology 15 (2024): 1393946. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0042] 42. Briand F., Mayoux E., Brousseau E., et al., “Empagliflozin, via Switching Metabolism Toward Lipid Utilization, Moderately Increases LDL Cholesterol Levels Through Reduced LDL Catabolism,” Diabetes 65 (2016): 2032–2038. [DOI] [PubMed] [Google Scholar]

[dom71390-bib-0043] 43. Bechmann L. E., Emanuelsson F., Nordestgaard B. G., and Benn M., “SGLT2‐Inhibition Increases Total, LDL, and HDL Cholesterol and Lowers Triglycerides: Meta‐Analyses of 60 Randomized Trials, Overall and by Dose, Ethnicity, and Drug Type,” Atherosclerosis 394 (2024): 117236. [DOI] [PubMed] [Google Scholar]

[dom71390-bib-0044] 44. Ferrannini E., Baldi S., Frascerra S., et al., “Shift to Fatty Substrate Utilization in Response to Sodium‐Glucose Cotransporter 2 Inhibition in Subjects Without Diabetes and Patients With Type 2 Diabetes,” Diabetes 65 (2016): 1190–1195. [DOI] [PubMed] [Google Scholar]

[dom71390-bib-0045] 45. Gübeli A., Steiner N., Limacher A., Mathis D., Melmer A., and Laimer M., “Dapagliflozin's Impact on Hormonal Regulation and Ketogenesis in Type 1 Diabetes: A Randomised Controlled Crossover Trial,” Diabetologia 68 (2025): 2116–2125. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0046] 46. Boeder S. C., Gregory J. M., Giovannetti E. R., and Pettus J. H., “SGLT2 Inhibition Increases Fasting Glucagon but Does Not Restore the Counterregulatory Hormone Response to Hypoglycemia in Participants With Type 1 Diabetes,” Diabetes 71 (2022): 511–519. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0047] 47. Martínez‐Reyes I. and Chandel N. S., “Mitochondrial TCA Cycle Metabolites Control Physiology and Disease,” Nature Communications 11 (2020): 102. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0048] 48. Mathew A. V., Jaiswal M., Ang L., Michailidis G., Pennathur S., and Pop‐Busui R., “Impaired Amino Acid and TCA Metabolism and Cardiovascular Autonomic Neuropathy Progression in Type 1 Diabetes,” Diabetes 68 (2019): 2035–2044. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0049] 49. Choi Y. J., Richard G., Zhang G., et al., “Attenuated Kidney Oxidative Metabolism in Young Adults With Type 1 Diabetes,” Journal of Clinical Investigation 134 (2024): e183984. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0050] 50. Kugathasan L., Sridhar V. S., Lovblom L. E., et al., “Interactive Effects of Empagliflozin and Hyperglycemia on Urinary Amino Acids in Individuals With Type 1 Diabetes,” Diabetes 73 (2024): 401–411. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0051] 51. Sakamoto M., Hasuzawa N., Wang L., et al., “Sodium‐Glucose Cotransporter 2 Inhibitors—But Not Insulin—Enhance Renal Branched‐Chain Amino Acid Catabolism,” Frontiers in Endocrinology 16 (2025): 1706838. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0052] 52. Nishitani S., Fukuhara A., Shin J., Okuno Y., Otsuki M., and Shimomura I., “Metabolomic and Microarray Analyses of Adipose Tissue of Dapagliflozin‐Treated Mice, and Effects of 3‐Hydroxybutyrate on Induction of Adiponectin in Adipocytes,” Scientific Reports 8 (2018): 8805. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0053] 53. Yamauchi Y., Nakamura A., Yokota T., et al., “Luseogliflozin Preserves the Pancreatic Beta‐Cell Mass and Function in db/db Mice by Improving Mitochondrial Function,” Scientific Reports 12 (2022): 9740. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0054] 54. Herat L. Y., Matthews J. R., Hibbs M., Rakoczy E. P., Schlaich M. P., and Matthews V. B., “SGLT1/2 Inhibition Improves Glycemic Control and Multi‐Organ Protection in Type 1 Diabetes,” iScience 26 (2023): 107260. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0055] 55. Hatanaka T., Ogawa D., Tachibana H., et al., “Inhibition of SGLT2 Alleviates Diabetic Nephropathy by Suppressing High Glucose‐Induced Oxidative Stress in Type 1 Diabetic Mice,” Pharmacology Research & Perspectives 4 (2016): e00239. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0056] 56. Wang S., Zhang Y., Balati A., Li B., Dai L., and Yu D., “Effect of Sodium‐Glucose Transporter 2 Inhibitors on Cardiovascular Outcomes in Type 1 Diabetes Mellitus: A Systematic Review and Meta‐Analysis of Randomised Controlled Trials,” Journal of Clinical Endocrinology and Metabolism 110, no. 7 (2025): 2071–2081, 10.1210/clinem/dgaf016. [DOI] [PubMed] [Google Scholar]

[dom71390-bib-0057] 57. Gessner A., Gemeinhardt A., Bosch A., et al., “Effects of Treatment With SGLT‐2 Inhibitors on Arginine‐Related Cardiovascular and Renal Biomarkers,” Cardiovascular Diabetology 21 (2022): 4. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0058] 58. Abhary S., Kasmeridis N., Burdon K. P., et al., “Diabetic Retinopathy Is Associated With Elevated Serum Asymmetric and Symmetric Dimethylarginines,” Diabetes Care 32 (2009): 2084–2086. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0059] 59. Omran D. R. A. R., Mohamed N. A. K., Ramzy T. R. A., and Ramiz S. A., “The Role of Serum Symmetric Dimethylarginine (SDMA) as Endogenous Biomarker of Renal Impairment in Correlation With Severity and Different Stages of Diabetic Kidney Disease,” Egyptian Journal of Internal Medicine 37 (2025): 165. [Google Scholar]

[dom71390-bib-0060] 60. Mishima E., Fukuda S., Kanemitsu Y., et al., “Canagliflozin Reduces Plasma Uremic Toxins and Alters the Intestinal Microbiota Composition in a Chronic Kidney Disease Mouse Model,” American Journal of Physiology. Renal Physiology 315 (2018): F824–F833. [DOI] [PubMed] [Google Scholar]

[dom71390-bib-0061] 61. Adingupu D. D., Göpel S. O., Grönros J., et al., “SGLT2 Inhibition With Empagliflozin Improves Coronary Microvascular Function and Cardiac Contractility in Prediabetic Ob/Ob−/− Mice,” Cardiovascular Diabetology 18 (2019): 16. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0062] 62. Herat L. Y., Magno A. L., Rudnicka C., et al., “SGLT2 Inhibitor–Induced Sympathoinhibition: A Novel Mechanism for Cardiorenal Protection,” JACC: Basic to Translational Science 5 (2020): 169–179. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0063] 63. Mutter S., Valo E., Aittomäki V., et al., “Urinary Metabolite Profiling and Risk of Progression of Diabetic Nephropathy in 2670 Individuals With Type 1 Diabetes,” Diabetologia 65 (2022): 140–149. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0064] 64. Niewczas M. A., Mathew A. V., Croall S., et al., “Circulating Modified Metabolites and a Risk of ESRD in Patients With Type 1 Diabetes and Chronic Kidney Disease,” Diabetes Care 40 (2017): 383–390. [DOI] [PMC free article] [PubMed] [Google Scholar]

[dom71390-bib-0065] 65. Lu Y.‐P., Zhang Z. Y., Wu H. W., et al., “SGLT2 Inhibitors Improve Kidney Function and Morphology by Regulating Renal Metabolic Reprogramming in Mice With Diabetic Kidney Disease,” Journal of Translational Medicine 20 (2022): 420. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Dapagliflozin Alters Plasma Metabolome in Patients With Type 1 Diabetes During Ketosis

Rosa Yang

Kevin Cho

Max C Petersen

Kai E Jones

Gary J Patti

Janet B McGill

Samantha A DiGruccio

ABSTRACT

Aims

Materials and Methods

Results

Conclusion

1. Introduction

2. Materials and Methods

2.1. Glucagon Analysis

2.2. Metabolomic Analysis, Relative Quantification

2.2.1. Polar Metabolites

2.2.2. Lipids

2.2.3. Ketone Body Derivatization

2.2.4. LC/MS Analysis

2.3. Statistical Analysis

3. Results

TABLE 1.

FIGURE 1.

FIGURE 2.

FIGURE 3.

FIGURE 4.

4. Discussion

Funding

Conflicts of Interest

Supporting information

Acknowledgements

Data Availability Statement

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases