Abstract
OBJECTIVE
Maturity-onset diabetes of the young (MODY) caused by pathogenic variants in HNF1A is a common form of monogenic diabetes. Sulfonylurea drugs are considered first-line treatment of HNF1A-MODY (MODY3), but intensified treatment is often needed. HNF1A encodes a transcription factor involved in the regulation of the sodium–glucose cotransporter 2 (SGLT2). Accordingly, the glucose-lowering efficacy of SGLT2 inhibitors in HNF1A-MODY is questionable. Here, we assess the glucose-lowering effect of the SGLT2 inhibitor empagliflozin as an add-on for treatment of individuals with HNF1A-MODY.
RESEARCH DESIGN AND METHODS
MOD3ST-TRIAL was a randomized, double-blind, placebo-controlled crossover trial. Adults with HNF1A-MODY treated with at least one glucose-lowering drug were randomized to be treated with empagliflozin 25 mg for 4 weeks followed by a 2-week washout period and then received placebo for 4 weeks or the opposite sequence. The primary outcome was mean glucose concentration assessed by 10 days of continuous glucose monitoring (CGM).
RESULTS
Nineteen individuals were randomized and 18 participants (n = 10 women [56%]; median [Q1, Q3] HbA1c 7.5% [7.0, 8.4] or 58 [53, 68] mmol/mol, mean [SD] CGM glucose concentration 10.4 [2.5] mmol/L) completed the study. Compared with placebo, empagliflozin lowered the mean glucose level 2.3 mmol/L (95% CI 1.3 to 3.3; P = 0.0001). There were no significant differences in hypoglycemic outcomes. Adverse events were generally mild and transient, and no severe adverse events or study drug discontinuations were attributable to empagliflozin.
CONCLUSIONS
Empagliflozin used for 4 weeks in adjunction with other glucose-lowering treatments markedly improved glycemia compared with placebo in individuals with HNF1A-MODY without significantly increasing risk of hypoglycemia or unexpected adverse effects.
Graphical Abstract
Introduction
Maturity-onset diabetes of the young (MODY) is a form of monogenic diabetes characterized by onset of hyperglycemia in early adulthood and autosomal dominant inheritance (1). Today, at least 10 genes are considered to cause a MODY phenotype (2,3). Pathogenic variants in the gene HNF1A cause a common MODY-subtype, HNF1A-MODY (also known as MODY3), which is characterized by impaired insulin secretion and normal insulin sensitivity (4,5). The risk of microvascular complications in individuals with HNF1A-MODY is comparable to that of individuals with type 2 diabetes (6,7). Accordingly, management of glycemia is important.
Currently, international guidelines recommend sulfonylureas (SUs) as first-line treatment, based on the results of a 6-week randomized crossover trial where an SU (gliclazide) was superior to metformin in lowering fasting plasma glucose levels in individuals with HNF1A-MODY (n = 16) (3,8). In a placebo-controlled, single-dose trial, nateglinide and glibenclamide (an SU) lowered plasma glucose levels during a meal test (n = 15) (9). Monotherapy with liraglutide was not superior to SU in a 6-week crossover trial (n = 16), although liraglutide was associated with glucose-lowering effects when compared with the study baseline when other glucose-lowering drugs were washed out (10). However, some individuals with HNF1A-MODY do not reach glycemic goals with SU monotherapy (11). Furthermore, treatment with SUs is associated with body weight gain and hypoglycemia (9,10,12,13). In a placebo-controlled crossover trial, the incretin-based treatment linagliptin used as an add-on to glimepiride had glucose-lowering effects in individuals with HNF1A-MODY (16 weeks; n = 19) (12).
To our knowledge, the use of sodium–glucose cotransporter 2 (SGLT2) inhibitors as a potential glucose-lowering strategy in individuals with HNF1A-MODY has not previously been investigated in a randomized controlled trial. SGLT2 inhibitors limit the reabsorption of glucose from the proximal tubules in the kidney and, consequently, increase urinary glucose excretion (14). In individuals with type 2 diabetes, SGLT2 inhibitors improve glycemic control without increasing the risk of hypoglycemia and body weight gain. Furthermore, SGLT2 inhibitors lower the risk of major cardiovascular events and chronic kidney disease (15,16). Importantly, HNF1A encodes a transcription factor that regulates expression of SGLT2 in the murine kidney and human renal cells (17,18). In humans with HNF1A-MODY, a lower renal glucose threshold and increased glucosuria have been attributed to a lower expression of SGLT2 (17,19). Yet, we have found, during a hyperglycemic clamp, that the effect of a single 25-mg dose of empagliflozin on urinary glucose excretion was not significantly different between individuals with type 2 diabetes and individuals with HNF1A-MODY (20). In case reports, SGLT2 inhibitors have been associated with improved glycemic control in individuals with HNF1A-MODY, with expected side effects, when used adjunctively to other glucose-lowering therapies (21,22).
The objective of the MOD3ST-TRIAL was to investigate the glucose-lowering effect of empagliflozin, 25 mg once daily, used adjunctively to other glucose-lowering treatments in individuals with HNF1A-MODY. We hypothesized that empagliflozin would lower plasma glucose concentration and improve glycemic control.
Research Design and Methods
Study Design
This study was an investigator-initiated, double-blind, placebo-controlled, randomized, superiority crossover trial in individuals with HNF1A-MODY conducted at two specialized diabetes clinics in Denmark (Steno Diabetes Center Copenhagen, Herlev; and Steno Diabetes Center Aarhus, Aarhus). The study was approved by the Danish Medical Research Ethics Committees and the Danish Medicines Agency (trial identifier EUCT 2023–503760-17-00) and conducted in accordance with Good Clinical Practice.
Participants
The target population were individuals with HNF1A-MODY who were treated with at least one glucose-lowering drug before inclusion. Inclusion criteria were 1) informed consent; 2) diabetes caused by a heterozygous likely pathogenic or pathogenic variant in HNF1A; 3) age 18 years or older; 4) glycated hemoglobin A1c (HbA1c) ≥6.5% (48 mmol/mol); 5) stable glucose-lowering treatment 60 days before inclusion; and 6) treatment with at least one glucose-lowering drug. Major exclusion criteria included estimated glomerular filtration rate (eGFR) <45 mL/min/1.73 m2, treatment with SGLT2 inhibitor 60 days before inclusion, and inability or unwillingness to abstain from a variable dosing regimen of glucose-lowering drugs (i.e., SUs, repaglinide, or insulin in a self-titrated regimen) during the study. The full list of eligibility criteria can be found in the Supplementary Appendix. Participants were asked for names of known first-degree relatives with HNF1A-MODY to assess relatedness between included participants.
Randomization
After the baseline period, participants were allocated to one of two intervention sequences: empagliflozin followed by placebo, or placebo followed by empagliflozin. A sequentially numbered randomization list was computer generated by the pharmacy of the Capital Region of Denmark using blocked randomization (block sizes of two and four, masked to blinded personnel). The study drug was packed by the pharmacy in containers labeled with randomization number and intervention period. Participants were allocated to intervention sequence at first drug dispensing. To blind the investigators and participants to allocation, empagliflozin (and a generic placebo tablet) was encapsulated in identical gelatin capsules (DBcaps, size AAel; Capsugel) by the pharmacy. Participants and investigators were blinded to the allocation of intervention sequence until after analysis of end points reported here.
Procedures
The study consisted of three periods: a baseline period (minimum 10 days) and two intervention periods (minimum 24 days each) separated by washout of the study drug (minimum 14 days) (Supplementary Fig. 1). Participants attended a total of seven visits (a visit was placed after at least 14 days of drug exposure in each intervention period).
The intervention consisted of the trial medication (orally administered 25 mg of empagliflozin or placebo, once daily) in addition to the participants’ existing glucose-lowering treatment. Existing glucose-lowering treatments were intended to be kept unchanged during the intervention unless changes were warranted due to hypoglycemia. Participants were encouraged to inform investigators about new-onset or assistance-requiring hypoglycemic events. If participant-reported hypoglycemia was considered problematic, hypoglycemic agents (i.e., SU, repaglinide, and insulins) were down-titrated at any time during the trial. For participants treated with glucose-lowering drugs in a variable-dosing regimen (i.e., SU, repaglinide, and insulins) at the screening visit, the variable-dosing regimen could be paused or adjusted to a stable dose regimen for the duration of the study if deemed unproblematic by the investigator.
During the baseline period and at the end of both intervention periods, blinded continuous glucose monitoring (CGM) was performed for 10 days using Dexcom G7 (a gift from Dexcom Inc., San Diego, CA). The sensor was inserted after the participant had been taking the study drug for at least 14 days. If a sensor failed, a new sensor was inserted, if possible, to obtain at least 10 days of sensor readings. Participants performed 24-h urine collection while CGM data were collected. Participants were instructed to measure the urine volume, take a sample of the 24-h urine collection, and keep the sample refrigerated until the next visit. On the last morning of the baseline period and of both intervention periods, the participants collected a fasting morning urine sample (to determine the urinary glucose-to-creatinine ratio) and measured a fasting plasma glucose concentration and a fasting plasma β-hydroxybutyrate (BHB) ketone concentration using a blood glucose/ketone meter (GlucoMen Areo 2 K; A. Menarini Diagnostics). Participants were given a study diary with instructions and forms for urine collection, self-measurement of fasting plasma concentrations of glucose and BHB and for recording of hypoglycemic events and unsolicited adverse events.
Compliance was assessed at visits by inquiry (“Do you still take the study drug?”) and by pill count at the end of each intervention period. Unsolicited adverse events were assessed by inquiry (“Have you had any problems since last time?”). Other safety procedures included blood pressure (measured in the office three times at every visit), and blood samples before and after each intervention to measure plasma sodium, potassium, creatinine, and alanine transferase.
Outcomes
The primary end point was the mean difference between empagliflozin and placebo in mean glucose level evaluated using CGM. The primary end point and secondary efficacy outcomes were analyzed in a principal stratum (efficacy population) excluding study participants with one or more of the following intercurrent events: 1) nonadherence to any of the assigned interventions (defined as participant-reported intentional discontinuation or pill count <75%) including discontinuation due to an adverse effect; and/or 2) initiation or dose change of glucose-lowering drugs during the study except for dose lowering of insulin, an SU, or repaglinide dose due to hypoglycemia. Secondary efficacy outcomes included CGM-derived metrics, 24-h urinary glucose excretion, estimated renal glucose threshold, fasting urinary glucose-to-creatine ratio, fasting plasma ketone (BHB) concentration, fasting plasma glucose concentration, and body weight. Secondary safety end points included CGM-derived and participant-reported hypoglycemic events and the number of participants with BHB concentrations ≥0.6 mmol/L and ≥1.5 mmol/L. Hypoglycemic events were assessed as binary outcomes (the number of participants with at least one event) and as rates (number of events) stratified by severity (level 1: plasma glucose concentration between 3.0 and 3.8 mmol/L; level 2: plasma glucose <3.0 mmol/L; and level 3: assistance-requiring hypoglycemia). Safety end points were evaluated under a “treatment policy strategy” including all participants who were randomized and received at least one dose of study drug regardless of compliance (safety population).
Laboratory Measurements
Urine was collected in containers without additives. Urine for analysis of glucose and creatinine was stored at −80°C until batch analysis (Cobas c 503; Roche Diagnostics). Other biochemical analyses were performed locally by an accredited hospital laboratory.
Calculations
All CGM readings were extracted using the manufacturer’s software and readings outside the limit of detection (<2.2 and >22.2 mmol/L) were imputed as 2.1 mmol/L and 22.3 mmol/L. Otherwise, no postprocessing of CGM readings was performed before calculation of CGM metrics. CGM metrics and hypoglycemic events were derived according to CGM consensus guidelines (23). Times in range were expressed as percentages and calculated as number of readings in the glycemic range divided by the total number of readings. A hypoglycemic event on CGM was defined as readings <3.9 mmol/L for ≥15 consecutive minutes, and the end of a hypoglycemic event was defined as readings of ≥3.9 mmol/L for 15 consecutive minutes (23). Hypoglycemic events including readings <3.0 mmol/L for ≥15 consecutive minutes were classified as level 2 and the remaining hypoglycemic events as level 1.
The renal glucose threshold was estimated using urinary glucose excretion, urinary creatinine clearance (a surrogate for GFR), and sensor glucose from CGM obtained during the 24-h urine collection according to a method described previously (24). Urinary creatinine clearance was calculated as urinary creatinine excretion rate divided by plasma creatinine concentration. If the urinary glucose excretion rate was <600 mg per 24 h, the renal glucose threshold was not estimated as prespecified (baseline: n = 4; placebo: n = 3) (24). As a sensitivity analysis, the 24-h renal glucose threshold was also estimated using GFR derived from plasma creatinine using the Chronic Kidney Disease Epidemiology Collaboration 2021 formula in combination with the Du Bois and Du Bois formula for body surface area (0.007184 × height [cm]0.725 × weight [kg]0.425).
HOMA2 for insulin resistance (HOMA2-IR) and β-cell function indices were calculated using HOMA calculator, version 2.2.3, software (Diabetes Trials Unit, University of Oxford, https://www.dtu.ox.ac.uk/homacalculator).
Statistical Analysis
To detect a target difference of 0.8 mmol/L in mean glucose (corresponding to an HbA1c difference of 5 mmol/mol or 0.5%) at a significance level of 5% with 80% power, at least 17 participants with complete data were required, assuming a between-subject SD of 1.7 mmol/L and an intrasubject correlation of 0.8 (implying the SD of the differences would be 1.1 mmol/L) (12,25,26). Hence, it was planned to continue recruitment until 18 participants had completed the study or 24 participants had been randomized to a study arm.
Continuous outcomes were evaluated in a linear mixed model with intervention and period as fixed effects and with an unstructured covariance to account for repeated measurements on each participant and potential variance heterogeneity between interventions. Degrees of freedom were computed using the Satterthwaite approximation. Binary safety outcomes were compared between interventions using the McNemar test, and event rates were compared between interventions with a negative binomial generalized linear mixed model with a random effect of study participant and an offset (duration of active CGM or length of intervention period). No tests of carryover effect were performed as recommended by International Council for Harmonisation E9 guidelines. Missing data in continuous outcomes and event rates were handled implicitly by maximum likelihood inference in the mixed models. The impact of missing values for binary outcomes was evaluated by using worst-case scenarios as a sensitivity analysis. A two-sided P value <0.05 was considered statistically significant for the primary outcome. P values for the secondary efficacy outcomes were adjusted for multiple testing using the method of Benjamini and Hochberg, which controls the false discovery rate, and an adjusted P value <0.05 was considered statistically significant. P values for safety end points were not adjusted for multiple testing.
Mixed-model analyses were performed with R statistical software, version 4.4.1; the LMMstar package, version 1.1.0; and the GLMMadaptive package, version 0.9–1. All statistical analyses were carried out in accordance with the protocol available from the EU Clinical Trials Information System (identifier EUCT 2023–503760-17-00), with minor corrections described in an addendum (https://osf.io/32vca) that was finalized before unblinding of the sequence allocation.
Data and Resource Availability
The data sets generated during and/or analyzed in this study are available from the corresponding author upon reasonable request and if in compliance with applicable data protection legislation. Study protocol and statistical analysis plan addendum are available at https://osf.io/u8ncg/.
Results
Participants were recruited between 11 October 2023 and 29 November 2024. Of 21 individuals assessed for eligibility, 19 individuals with HNF1A-MODY were eligible, randomized to a treatment sequence, and included in the safety analysis. One participant (allocated to placebo first) withdrew from the trial 2 weeks after randomization in relation to worsening of an existing condition before any outcome measures were assessed. Eighteen participants randomized equally in number to both sequences (both arms, n = 9) completed the trial without occurrence of any of the prespecified intercurrent events (nonadherence to study drug or change of concurrent glucose-lowering therapy due to other factors than hypoglycemia) and were included in the efficacy analysis (Supplementary Fig. 2). Compliance assessed as pill count was high during both interventions (placebo: ≥92%, maximum of two missed capsules; empagliflozin: ≥93%, maximum of two missed capsules).
The participant characteristics for the efficacy population (n = 18) are presented in Table 1. The participants carried the following HNF1A variants: p.Leu30Pro (n = 1), p.Asp135Asn (n = 1), exon 2–10 deletion (n = 2), p.Gly288fs (n = 1), p.Pro289fs (n = 1), p.Pro291fs (n = 7), c.956-2A>G (n = 2), p.Glu332Ter (n = 1), p.Pro379fs (n = 1), and p.Val590fs (n = 1). Four participants were first-degree relatives (two siblings in one family, a mother and daughter in another family). All participants self-identified as Danish regarding ethnicity.
Table 1.
Participant characteristics
| Characteristic | Data |
|---|---|
| Total patients, N | 18 |
| Men, n (%) | 8 (44) |
| Age, years | 48 [39, 55] |
| Onset age, years | 20 [15, 27] |
| Diabetes duration, years | 26 [20, 30] |
| Body weight, kg | 77 (12) |
| BMI, kg/m2 | 24.8 (2.8) |
| eGFR, mL/min/1.73 m2 | 110 [95, 116] |
| HbA1c, % | 7.5 [7.0, 8.4] |
| HbA1c, mmol/mol | 58 [53, 68] |
| Total cholesterol, mmol/L | 4.4 [4.0, 5.3] |
| LDL cholesterol, mmol/L | 2.4 [2.1, 3.0] |
| HDL cholesterol, mmol/L | 1.4 [1.3, 1.7] |
| Urine albumin-to-creatinine ratio, mg/g | 5 [3, 8] |
| Fasting venous plasma glucose, mmol/L | 8.8 [7.9, 10.9] |
| Fasting C-peptide, pmol/L | 479 [314, 711] |
| Fasting insulin, pmol/L | 38 [21, 53] |
| HOMA2 insulin resistance | 1.28 [1.03, 1.92], n = 15 |
| HOMA2 β-cell function, % | 37 [24, 48], n = 15 |
| Complications, n (%) | |
| Retinopathy | 7 (39) |
| Nephropathy | 2 (11) |
| Neuropathy | 2 (11) |
| Cerebrovascular disease | 0 (0) |
| Coronary heart disease | 1 (6) |
| Peripheral arterial disease | 1 (6) |
| Glucose-lowering drugs, n (%) | |
| SU | 16 (89) |
| GLP-1RA | 3 (17) |
| Metformin | 3 (17) |
| Repaglinide | 1 (6) |
| DPP-4 inhibitor | 6 (33) |
| Insulin, bolus | 2 (11) |
| Insulin, basal | 3 (17) |
Data are reported as mean (SD) or median [quartile 1, quartile 3]. HOMA2 measures are only reported for participants not treated with insulin (n = 15). DPP-4, dipeptidyl peptidase-4; GLP-1RA, glucagon-like peptide-1 receptor agonist.
Empagliflozin significantly lowered mean glucose evaluated by CGM (primary outcome) and time above range compared with placebo (Table 2 and Fig. 1). Also, time in range and time in tight range increased with empagliflozin. Furthermore, empagliflozin lowered fasting plasma glucose levels. Glycemic variability as measured by the SD was significantly lower with empagliflozin compared with placebo but did not differ significantly when measured by the coefficient of variation, another measure of glycemic variability. No significant differences between empagliflozin and placebo were observed for time below range (<3.0 mmol/L and <3.9 mmol/L, respectively) (Table 2).
Table 2.
Efficacy outcomes
| Outcome | Baseline | Placebo | Empagliflozin | ETD (95% CI) | P value | FDR |
|---|---|---|---|---|---|---|
| Mean glucose, mmol/L | 10.4 (2.5) | 10.2 (2.2) | 7.9 (1.5) | −2.3 (−3.3 to −1.3) | 0.0001 | — |
| CV, % | 29.6 (6.9) | 30.2 (7.3) | 32.2 (5.6) | 2.0 (−1.3 to 5.3) | 0.21 | 0.27 |
| SD, mmol/L | 3.0 (0.8) | 3.0 (0.8) | 2.6 (0.9) | −0.5 (−0.7 to −0.2) | 0.0029 | 0.0041 |
| Time in range, % | 52 (26) | 52 (22) | 78 (14) | 26 (16 to 37) | 0.0001 | 0.0003 |
| Time in tight range, % | 28 (20) | 28 (18) | 56 (14) | 28 (18 to 38) | <0.0001 | <0.0001 |
| Time above range, % | 46 (26) | 47 (24) | 19 (14) | −27 (−39 to −16) | 0.0002 | 0.0004 |
| Time below range (<3.9 mmol/L), % | 0.0 [0.0, 0.6] | 0.1 [0.0, 1.5] | 1.3 [0.2, 4.3] | 1.0 (−0.7 to 2.7) | 0.23 | 0.27 |
| Time below range (<3.0 mmol/L), % | 0.0 [0.0, 0.0] | 0.0 [0.0, 0.1] | 0.2 [0.0, 0.5] | 0.2 (−0.4 to 0.8) | 0.44 | 0.44 |
| 24-h urinary glucose excretion, g | 5 [0, 26] | 4 [1, 10] | 102 [73, 123] | 77 (48 to 106) | <0.0001 | 0.0001 |
| Missing data | 1 | 1 | 0 | |||
| Renal glucose threshold (CC-based), mmol/L | 12.3 (2.5) | 12.3 (1.6) | 4.4 (0.9) | −7.7 (−9.1 to −6.4) | <0.0001 | <0.0001 |
| Missing data | 7 | 5 | 4 | |||
| Renal glucose threshold (eGFR-based), mmol/L | 12.0 (2.4) | 12.0 (1.6) | 4.0 (1.0) | −8.2 (−9.2 to −7.2) | <0.0001 | — |
| Missing data | 7 | 5 | 3 | |||
| Urinary glucose-to-creatinine ratio, mol/mol | 1.2 [0.0, 5.4] | 0.2 [0.0, 1.1] | 29.7 [23.7, 41.4] | 32.1 (25.4 to 38.9) | <0.0001 | <0.0001 |
| Missing data | 1 | 1 | 1 | |||
| Fasting plasma ketone (BHB), mmol/L | 0.3 (0.2) | 0.4 (0.3) | 0.3 (0.2) | −0.1 (−0.2 to 0.1) | 0.31 | 0.34 |
| Fasting plasma glucose, mmol/L | 8.1 (2.2) | 8.0 (2.2) | 5.9 (0.9) | −2.1 (−3.2 to −1.1) | 0.0005 | 0.0009 |
| Body weight, kg | 77.0 (12.5) | 77.0 (12.6) | 76.3 (12.2) | −0.7 (−1.1 to −0.3) | 0.0016 | 0.0026 |
Efficacy outcomes in principal stratum population (n = 18). Data are mean (SD) or median [quartile 1, quartile 3] unless otherwise indicated. Reasons for missing data are described in Supplementary Table 2. CC, creatinine clearance (used as glomerular filtration rate in estimation of renal glucose threshold); CV, coefficient of variation (CGM metric); eGFR, estimated glomerular filtration rate (used as glomerular filtration rate in estimation); ETD, estimated treatment difference; FDR, false discovery rate.
Figure 1.
CGM metrics, mean glucose (A) and time in ranges (B) (level 2, >13.9 mmol/L; level 1, 10.1–13.9 mmol/L; time in range, 3.9–10.0 mmol/L; and time below range, <3.9 mmol/L). Data are means and SD.
Empagliflozin led to marked increases in fasting urinary glucose-to-creatinine ratio, 24-h urinary glucose excretion, and lowering of the renal glucose threshold compared with placebo. A small but statistically significant reduction in body weight was observed after treatment with empagliflozin compared with placebo (Table 2). No difference between interventions was observed in terms of fasting plasma BHB ketone levels (Table 2). No fasting BHB measurements were ≥1.5 mmol/L. Two participants had a fasting BHB concentration ≥0.6 mmol/L (one participant during placebo and one participant during empagliflozin; P > 0.99).
The number of participants with hypoglycemia and the number of events per participant are summarized in Table 3. There were no severe (level 3) hypoglycemic events. Overall, there were no significant differences between interventions for the rates of hypoglycemic events or the number of participants with hypoglycemic events stratified by severity levels. The participant who discontinued the study did not report any hypoglycemic events. At the end of the baseline period, the SU dose was reduced in two participants due to hypoglycemia according to CGM. In the second intervention period, participant-reported hypoglycemia led to down-titration of SU in one participant while receiving placebo and insulin in another participant while receiving empagliflozin.
Table 3.
Hypoglycemic events
| Level | Placebo | Empagliflozin | Estimated rate ratio† (95% CI) | Participants with events during both interventions, n‡ | ||||
|---|---|---|---|---|---|---|---|---|
| Participants with events, n/n (%) | Total events | Events per participant, n* | Participants with events, n/n (%) | Total events | Events per participant, n* | |||
| CGM | ||||||||
| 1 + 2 | 8/18 (44) | 42 | 0 [0, 2] | 12/18 (67) | 72 | 4 [0, 6] | 2.5 (0.7 to 8.6), P = 0.15 | 7, P = 0.22 (0.13) |
| 1§ | 8/18 (44) | 31 | 0 [0, 2] | 12/18 (67) | 57 | 3 [0, 5] | 1.9 (0.6 to 5.9), P = 0.27 | 7, P = 0.22 (0.13) |
| 2‖ | 4/18 (22) | 11 | 0 [0, 0] | 8/18 (44) | 15 | 0 [0, 2] | 1.9 (0.6 to 6.1), P = 0.30 | 2, P = 0.29 (0.18) |
| Participant-reported events | ||||||||
| 1 + 2 | 3/19 (16) | 7 | 0 [0, 0] | 5/18 (28) | 9 | 0 [0, 1] | 1.3 (0.4 to 4.2), P = 0.64 | 3, P = 0.48 (0.25) |
| 1 | 3/19 (16) | 5 | 0 [0, 0] | 5/18 (28) | 7 | 0 [0, 1] | 1.4 (0.4 to 4.7), P = 0.59 | 3, P = 0.48 (0.25) |
| 2 | 2/19 (11) | 2 | 0 [0, 0] | 2/18 (11) | 2 | 0 [0, 0] | — | 1, P > 0.99 (>0.99) |
Hypoglycemic events in the safety population (n = 19). No level 3 hypoglycemic events (i.e., severe events requiring assistance) occurred.
*Events per participant are given as the observed count (median [quartile 1, quartile 3]).
†Rate ratio could not be estimated for participant-reported hypoglycemic events level 2 owing to convergence issues.
‡Significance test for the number of participants with event(s) derived from the McNemar test (in parentheses: worst-case analysis assuming a hypoglycemic event during empagliflozin for participant who discontinued the study).
§Level 1 hypoglycemia: plasma glucose between 3.0 and 3.8 mmol/L.
‖Level 2 hypoglycemia: plasma glucose <3.0 mmol/L.
Between randomization and the last visit, 18 adverse events occurred (excluding participant-reported hypoglycemic events: n = 6 events in participants receiving placebo; n = 9 events while receiving empagliflozin; and n = 3 events during washout). One serious adverse event occurred in the washout period after placebo (hospitalization due to renal calculi). The adverse events were generally mild and transient (Supplementary Table 1) and included side effects known to SGLT2 inhibitors (empagliflozin: balanitis [n = 2 of 8 men], vulvovaginitis [n = 1 of 10 women], urinary tract infection [n = 1 of 18 participants]; placebo: urinary tract infection [n = 1 of 19 participants]). The participant who discontinued the study had two adverse events during placebo (rash and upper respiratory infection).
Conclusions
In individuals with HNF1A-MODY, 4 weeks of treatment with the SGLT2 inhibitor empagliflozin significantly lowered mean glucose assessed by CGM compared with placebo when added to existing background treatment. Correspondingly, other measures of glycemic control, including fasting plasma glucose level, improved. Furthermore, no significant difference in measures of hypoglycemia were found.
Strengths of the study include the randomized, double-blind, placebo-controlled design. A crossover design was chosen to decrease the sample size. The pharmacokinetics of empagliflozin (half-life ∼13 h) and the rapid onset of glucose-lowering effects of SGLT2 inhibitors allow investigation of empagliflozin in studies with relative short duration. Based on the pharmacokinetics of empagliflozin, 2 weeks was considered adequate for washout and for steady state at the measurement of the primary outcome. Carryover effects are theoretically implausible, and concerns of carryover effects of SGLT2 inhibitors in crossover studies have not been raised, to our knowledge.
The efficacy end points were evaluated in principal stratum consisting of participants who tolerated both placebo and empagliflozin and adhered to treatment. This approach was chosen because it answers questions relevant for clinicians and patients (“what is the effect if the drug is used as intended”) and reduces variation due to nonadherence under a “treatment policy strategy” (intention-to-treat analysis). However, only one participant discontinued the study after exposure to placebo only and was excluded from the efficacy analysis. Thus, the results are comparable to what would have been obtained with a treatment policy strategy. Furthermore, safety outcomes were evaluated in all randomized participants regardless of adherence.
The short study duration and small sample size are comparable to previous trials of treatments in individuals with HNF1A-MODY but nevertheless limit comprehensive assessment of safety and long-term efficacy. The rates of hypoglycemic events were low and the finding that rates of hypoglycemia were not statistically significant between interventions may be a type II error (false negative). Furthermore, the study was not designed as a treat-to-target study with titration of insulin and SU. Thus, the possibility of preventing hypoglycemia by adjustment of glucose-lowering agents associated with hypoglycemia (i.e., insulins and SUs) was not investigated in this study.
The eligibility criteria did not target a specific population with regard to use of concurrent glucose-lowering drugs to ease recruitment. Participants should be treated with at least one other glucose-lowering drug prior to and during the study. This is comparable to the treatment algorithm used in type 2 diabetes where SGLT2 inhibitors are initiated in individuals with type 2 diabetes adjunctively to a diverse range of other glucose-lowering drugs. Most participants were treated with an SU (89%); the use of other glucose-lowering drugs was heterogenous and included incretin-based drugs (50%), metformin (17%), and insulins (17%). The heterogenous use of concurrent glucose-lowering treatments limits the conclusions on how SGLT2 inhibitors could be implemented in treatment guidelines relative to other glucose-lowering drugs. However, HNF1A-MODY is characterized by an impaired insulin secretion that can often be ameliorated by SUs and incretin-based therapies. Theoretically, improving the insulin secretion with SU and β-cell sensitivity to glucose with GLP-1–based therapy and then adding an SGLT2 inhibitor constitute a treatment strategy supported by mechanistic rationale (8,10,12,27). Due to the small sample size, subgroup analysis was futile. Importantly, the trial did not evaluate empagliflozin as a monotherapy, and we cannot draw conclusions about the use of SGLT2 inhibitors as first-line treatment.
To estimate the renal glucose threshold, the 24-h urine collection had to be collected when the CGM was active. This implies that the 24-h urine samples could not be frozen immediately after collection but were stored refrigerated by the participants until the next visit (the median time from end of collection to visit was 4 days). Urine glucose may have been degraded, which may have resulted in overestimation of the renal glucose threshold. Furthermore, the estimation of the renal glucose threshold used is based on an idealized threshold, which theoretically overestimates the actual renal glucose threshold (14).
This randomized clinical trial appears to be the first to assess the effect of an SGLT2 inhibitor on glycemic control in individuals with monogenic diabetes. The role of SGLT2 inhibitors in the treatment of individuals with HNF1A-MODY has been questioned due to the suggested transcriptional regulation of SLC5A2 (the gene encoding SGLT2) by the transcription factor HNF1A (17,18). Two studies found a markedly lower renal glucose threshold in individuals with HNF1A-MODY compared with individuals with type 1 diabetes or type 2 diabetes, respectively (17,19). The quantitative impact of the lowered renal glucose threshold on the urinary glucose excretion has not been assessed until recently; we measured the urinary glucose excretion during a hyperglycemic clamp with and without SGLT2 inhibition (20). In comparison with individuals with type 2 diabetes, the urinary glucose excretion was not significantly decreased in individuals with HNF1A-MODY, and the effect of SGLT2 inhibition was comparable between the two groups. In the phase 3 program for empagliflozin in individuals with type 2 diabetes, empagliflozin reduced HbA1c with ∼0.7% (7 mmol/mol; equivalent to CGM mean glucose ∼1.5 mmol/L) after 24 weeks of treatment from baseline HbA1c ∼8.0% (64 mmol/mol; equivalent to CGM mean glucose of ∼11 mmol/L) (28). Thus, the glucose-lowering effect of empagliflozin in individuals with HNF1A-MODY after 4 weeks of treatment seems to be comparable to that observed in type 2 diabetes after 24 weeks, but our study was too short to assess effects on HbA1c. Empagliflozin reduces the risk of major cardiovascular events and chronic kidney disease in individuals with type 2 diabetes (15,16). Individuals with HNF1A-MODY, who are susceptible to both microvascular and macrovascular complications, may potentially benefit from these protective effects (6,7,29). However, longer and larger trials with other primary end points are required to clarify the cardio- and nephroprotective effects in individuals with HNF1A-MODY.
Although the number of hypoglycemic events was greater during empagliflozin treatment, especially for hypoglycemic events on CGM, no statistically significant differences between interventions were observed. The number of level 2 hypoglycemic events (<3.0 mmol/L) reported by participants was equal between interventions, and time below range (<3.0 mmol/L) was 8 min/day in most participants (75%) during empagliflozin treatment. However, SGLT2 inhibitors do increase the risk of (mild) hypoglycemia when added to treatment with an SU in individuals with type 2 diabetes (30). Because SUs are considered first-line treatment of individuals with HNF1A-MODY, the risk of hypoglycemia should be addressed when initiating an SGLT2 inhibitor in a person treated with insulin and/or an SU. Future, longer clinical studies of SGLT2 inhibitors in HNF1A-MODY should consider addressing prevention of hypoglycemia by down-titration of SU and/or insulin eventually combined with incretin-based treatments.
Diabetic ketoacidosis due to SGLT2 inhibitors is a rare side effect observed in individuals with type 2 diabetes or with type 1 diabetes (30–32), and the present study was too short and small to assess this risk. In a case report of a person with HNF1A-MODY treated with dapagliflozin as monotherapy and a very-low-carbohydrate diet, ketonemia (>3.0 mmol/L, without acidosis) led to discontinuation of dapagliflozin (33). In the present study, levels of a fasting plasma ketone (BHB) were not affected by empagliflozin; the participants had residual β-cell function and were treated with other glucose-lowering drugs that stimulate insulin secretion (Table 1). Precautions to avoid diabetic ketoacidosis are potentially the same as for other forms of diabetes (e.g., interruption of treatment during major surgery or acute illness; focus on potential precipitating factors such as low β-cell function, reduced insulin doses, increased insulin requirements or restricted food intake) (32). In individuals with HNF1A-MODY, low β-cell function could be indicated by a poor glycemic response to an SU or low plasma C-peptide level. In these cases, incretin-based treatments (together with an SU) may improve β-cell function and thereby constitute an alternative or complementary treatment to SGLT2 inhibitors (10,12,21); however, whether combination with incretin-based therapy prevents diabetic ketoacidosis is highly speculative. In this study, adverse events related to empagliflozin were mild and/or expected and did not lead to discontinuation; observed known side effects of SGLT2 inhibitors were balanitis, vulvovaginitis, and urinary tract infection. Information to users about SGLT2-induced genital infections could include how to mitigate and handle these, because this might affect treatment adherence.
In this randomized, placebo-controlled trial conducted with individuals with HNF1A-MODY, empagliflozin as an add-on therapy improved glycemic control evaluated by CGM metrics and fasting plasma glucose without significantly increasing the risk of hypoglycemia or other adverse events. Together with published case reports, this study supports that SGLT2 inhibitors have a clinically relevant glucose-lowering effect in individuals with HNF1A-MODY despite the potentially reduced expression of SGLT2 and that empagliflozin could be used as a second-line and/or third-line agent.
This article contains supplementary material online at https://doi.org/10.2337/figshare.30874850.
Article Information
Acknowledgments. The authors thank Alexanders S. Christensen (Roskilde, Denmark) for his invaluable contribution in drafting an initial synopsis for the study design.
No funder had any role in study design, data collection or interpretation, or writing of the manuscript.
Duality of Interest. F.K.K. has served on scientific advisory panels and/or been part of speakers bureaus for, owns stocks in, served as a consultant to, and/or received research support from 89bio, Amgen, AstraZeneca, Boehringer Ingelheim, Carmot Therapeutics, Eli Lilly, Gubra, MedImmune, MSD/Merck, Mundipharma, Norgine, Novo Nordisk, Sanofi, Structure Therapeutics, Zealand Pharma and Zucara; he is a co-founder of and minority shareholder in Antag Therapeutics; and has been employed at Novo Nordisk A/S since December 2023. T.H. owns stocks in Novo Nordisk and has received research support from Novo Nordisk and GSK. S.H. has served as a consultant to Novo Nordisk, served on scientific advisory board for Boehringer Ingelheim, and received travel reimbursement for conference attendance from Novo Nordisk. T.V. has served on scientific advisory panels and/or speakers bureaus or has served as a consultant to and/or received research support from Amgen, AstraZeneca, Bristol Myers Squibb, Boehringer Ingelheim, Carmot Therapeutics, Eli Lilly, GSK, Novo Nordisk, Zealand Pharma, Roche, Sanofi, and Sun Pharmaceuticals. No other potential conflicts of interest relevant to this article were reported.
Author Contributions. H.M., F.K.K., T.H., S.H., and T.V. conceptualized and designed the trial. J.S. and T.V. were principal investigators at each study site. H.M., S.O.H., J.S.J., and J.S. contributed to data acquisition. H.M., H.H.T., H.V., J.S., and T.V. recruited the participants. A.C.B.T. and T.H. contributed to the interpretation of genetic data. H.M. and J.L.F. contributed to the statistical analysis. All authors critically edited the manuscript and approved the final version. H.M. and T.V. are the guarantors of this work and, as such, had full access to all the data in the study and take 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 61st Annual Meeting of the European Association for the Study of Diabetes, Vienna, Austria, 16–19 September 2025.
Handling Editors. The journal editors responsible for overseeing the review of the manuscript were John B. Buse and Vanita R. Aroda.
Funding Statement
The study was indirectly supported by the Novo Nordisk Foundation through unrestricted grants to the Steno Diabetes Center Copenhagen. Dexcom Inc. provided the CGM sensors free of charge (without any influence on the study protocol and conduct).
Footnotes
Clinical trial reg. no. EUCT 2023-503760-17-00, euclinicaltrials.eu
Contributor Information
Henrik Maagensen, Email: henrik.maagensen@regionh.dk.
Tina Vilsbøll, Email: tina.vilsboell.01@regionh.dk.
Supporting information
References
- 1. Tattersall RB, Fajans SS.. A difference between the inheritance of classical juvenile-onset and maturity-onset type diabetes of young people. Diabetes 1975;24:44–53 [DOI] [PubMed] [Google Scholar]
- 2. Billings LK, Shi Z, Resurreccion WK, et al. Statistical evidence for high-penetrance MODY-causing genes in a large population-based cohort. Endocrinol Diabetes Metab 2022;5:e372. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Greeley SAW, Polak M, Njølstad PR, et al. ISPAD Clinical Practice Consensus Guidelines 2022: the diagnosis and management of monogenic diabetes in children and adolescents. Pediatr Diabetes 2022;23:1188–1211 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Pearson ER, Velho G, Clark P, et al. β-Cell genes and diabetes: quantitative and qualitative differences in the pathophysiology of hepatic nuclear factor-1α and glucokinase mutations. Diabetes 2001;50(Suppl. 1):S101–S107 [DOI] [PubMed] [Google Scholar]
- 5. Tripathy D, Carlsson AL, Lehto M, et al. Insulin secretion and insulin sensitivity in diabetic subgroups: studies in the prediabetic and diabetic state. Diabetologia 2000;43:1476–1483 [DOI] [PubMed] [Google Scholar]
- 6. Velho G, Vaxillaire M, Boccio V, et al. Diabetes complications in NIDDM kindreds linked to the MODY3 locus cn Chromosome 12q. Diabetes Care 1996;19:915–919 [DOI] [PubMed] [Google Scholar]
- 7. Isomaa B, Henricsson M, Lehto M, et al. Chronic diabetic complications in patients with MODY3 diabetes. Diabetologia 1998;41:467–473 [DOI] [PubMed] [Google Scholar]
- 8. Pearson ER, Starkey BJ, Powell RJ, et al. Genetic cause of hyperglycaemia and response to treatment in diabetes. Lancet 2003;362:1275–1281 [DOI] [PubMed] [Google Scholar]
- 9. Tuomi T, Honkanen EH, Isomaa B, et al. Improved prandial glucose control with lower risk of hypoglycemia with nateglinide than with glibenclamide in patients with maturity-onset diabetes of the young type 3. Diabetes Care 2006;29:189–194 [DOI] [PubMed] [Google Scholar]
- 10. Østoft SH, Bagger JI, Hansen T, et al. Glucose-lowering effects and low risk of hypoglycemia in patients with maturity-onset diabetes of the young when treated with a GLP-1 receptor agonist: a double-blind, randomized, crossover trial. Diabetes Care 2014;37:1797–1805 [DOI] [PubMed] [Google Scholar]
- 11. Shepherd MH, Shields BM, Hudson M, et al.; UNITED study . A UK nationwide prospective study of treatment change in MODY: genetic subtype and clinical characteristics predict optimal glycaemic control after discontinuing insulin and metformin. Diabetologia 2018;61:2520–2527 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Christensen AS, Hædersdal S, Støy J, et al. Efficacy and safety of glimepiride with or without linagliptin treatment in patients with HNF1A diabetes (maturity-onset diabetes of the young type 3): a randomized, double-blinded, placebo-controlled, crossover trial (GLIMLINA). Diabetes Care 2020;43:2025–2033 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. UK Prospective Diabetes Study (UKPDS) Group . Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet 1998;352:837–853 [PubMed] [Google Scholar]
- 14. Al-Jobori H, Daniele G, Cersosimo E, et al. Empagliflozin and kinetics of renal glucose transport in healthy individuals and individuals with type 2 diabetes. Diabetes 2017;66:1999–2006 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Herrington WG, Staplin N, Wanner C, et al.; EMPA-KIDNEY Collaborative Group . Empagliflozin in patients with chronic kidney disease. N Engl J Med 2023;388:117–127 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Zinman B, Wanner C, Lachin JM, et al.; EMPA-REG OUTCOME Investigators . Empagliflozin, cardiovascular outcomes, and mortality in type 2 diabetes. N Engl J Med 2015;373:2117–2128 [DOI] [PubMed] [Google Scholar]
- 17. Pontoglio M, Prié D, Cheret C, et al. HNF1alpha controls renal glucose reabsorption in mouse and man. EMBO Rep 2000;1:359–365 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Takesue H, Hirota T, Tachimura M, et al. Nucleosome positioning and gene regulation of the SGLT2 gene in the renal proximal tubular epithelial cells. Mol Pharmacol 2018;94:953–962 [DOI] [PubMed] [Google Scholar]
- 19. Menzel R, Kaisaki PJ, Rjasanowski I, et al. A low renal threshold for glucose in diabetic patients with a mutation in the hepatocyte nuclear factor-1α (HNF-1α) gene. Diabet Med 1998;15:816–820 [DOI] [PubMed] [Google Scholar]
- 20. Maagensen H, Jensen JS, Høyerup SO, et al. Effect of SGLT2 inhibition on glucosuria during a hyperglycemic clamp in HNF1A-MODY (MODY3) and type 2 diabetes. Diabetes Care 2025;48:1536–1544 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Maagensen H, Hædersdal S, Krogh J, et al. Clinical effects of SGLT2 inhibitors in seven persons with HNF1A-MODY (MODY3). Diabet Med 2024;41:e15387. [DOI] [PubMed] [Google Scholar]
- 22. Sriravindrarajah A, Fernandes A, Wu T, et al. The use of SGLT2 inhibitors in achieving glycaemic control in maturity-onset diabetes of the young type 3. Endocrinol Diabetes Metab Case Rep 2021;2021:21–0102 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Battelino T, Alexander CM, Amiel SA, et al. Continuous glucose monitoring and metrics for clinical trials: an international consensus statement. Lancet Diabetes Endocrinol 2023;11:42–57 [DOI] [PubMed] [Google Scholar]
- 24. Polidori D, Sha S, Ghosh A, et al. Validation of a novel method for determining the renal threshold for glucose excretion in untreated and canagliflozin-treated subjects with type 2 diabetes mellitus. J Clin Endocrinol Metab 2013;98:E867–71 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Riddlesworth TD, Beck RW, Gal RL, et al. Optimal sampling duration for continuous glucose monitoring to determine long-term glycemic control. Diabetes Technol Ther 2018;20:314–316 [DOI] [PubMed] [Google Scholar]
- 26. Selvin E, Wang D, Rooney MR, et al. Within-person and between-sensor variability in continuous glucose monitoring metrics. Clin Chem 2023;69:180–188 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Christensen AS, Hædersdal S, Storgaard H, et al. GIP and GLP-1 potentiate sulfonylurea-induced insulin secretion in hepatocyte nuclear factor 1α mutation carriers. Diabetes 2020;69:1989–2002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Levine MJ. Empagliflozin for type 2 diabetes mellitus: an overview of phase 3 clinical trials. Curr Diabetes Rev 2017;13:405–423 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Steele AM, Shields BM, Shepherd M, et al. Increased all-cause and cardiovascular mortality in monogenic diabetes as a result of mutations in the HNF1A gene. Diabet Med 2010;27:157–161 [DOI] [PubMed] [Google Scholar]
- 30. Kinduryte Schorling O, Clark D, Zwiener I, et al. Pooled safety and tolerability analysis of empagliflozin in patients with type 2 diabetes mellitus. Adv Ther 2020;37:3463–3484 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Rosenstock J, Marquard J, Laffel LM, et al. Empagliflozin as adjunctive to insulin therapy in type 1 diabetes: the EASE Trials. Diabetes Care 2018;41:2560–2569 [DOI] [PubMed] [Google Scholar]
- 32. Jardiance - Summary of Product Characteristics (SmPC). European Medicines Agency, 2025. Accessed 26 November 2025. Available from https://www.ema.europa.eu/en/medicines/human/EPAR/jardiance [Google Scholar]
- 33. Phan F, Bourron O, Laroche S, et al. Euglycaemic diabetic ketosis decompensation under dapagliflozin in a patient with MODY3. Diabetes Metab 2021;47:101248. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.