Eight weeks of empagliflozin does not affect pancreatic fat content and insulin secretion in people with prediabetes

Abstract

Fat accumulation within the pancreatic parenchyma is known to be a risk factor for a number of conditions, including pancreatic cancer and chronic pancreatitis.1 There is also growing evidence that the accumulation of pancreatic fat may contribute to the pathogenesis of diabetes.1, 2 Adipocytes within the pancreas could impair insulin secretion via secreted factors.1, 3 Some studies reported that fatty pancreas is linked to reduced insulin secretion, particularly in persons with prediabetes, while other trials did not detect any such associations.1 Previous findings of the DiRECT lifestyle intervention trial emphasize that pancreatic fat reduction plays a crucial role in the improvement of β-cell function and remission of type 2 diabetes by hypocaloric diet.4 Pharmacological approaches for reducing fat in the pancreas are, however, still lacking. The substance class of sodium-glucose cotransporter 2 (SGLT2) inhibitors can lower ectopic fat accumulation in the liver and shift energy metabolism towards fat oxidation.5 We have now tested the ability of empagliflozin to reduce pancreatic fat content in overweight and obese persons with prediabetes. The study protocol of this double-blind, placebo-controlled randomized trial was approved by the local ethics committee. Participants provided informed written consent before enrolment. In total, 40 participants (24 women, 16 men, 60 ± 9 years, body mass index 31.5 ± 3.8 kg/m2; mean ± SD) with prediabetes [according to the American Diabetes Association oral glucose tolerance test (OGTT) criteria] were randomized 1:1, receiving 25 mg empagliflozin daily or placebo for 8 weeks (Figure 1A). Thereof, 15 showed impaired fasting glucose, seven impaired glucose tolerance and 18 both. Detailed methods and patient characteristics were reported previously together with the primary endpoint.6 In short, we observed improved insulin sensitivity of the hypothalamus in response to empagliflozin. This was accompanied by lowering of fasting glucose and reduction of liver fat.6 Before and after treatment, pancreatic fat accumulation was quantified by magnetic resonance imaging (3T Magnetom Vida; Siemens Healthcare, Erlangen, Germany) following an overnight fast (Figure 1B). An 18-channel body-array receiver coil was placed on the upper abdomen of the participants, which were placed in a supine position. Localizer imaging was subsequently performed by applying a single breath-hold gradient echo imaging technique in all three spatial directions (10 s) for planning of the chemical shift encoded Dixon sequence, enabling quantification of proton density fat fraction (PDFF). A 3D multi-echo CSE sequence covering the abdominal organs, entire liver and pancreas was applied. Detailed sequence parameters: matrix size 160 × 132, field-of-view 380 × 314 mm, partition thickness 3 mm with 80 partitions in total. Repetition time was 8.9 ms, six echoes with echo times were 1.09, 2.46, 3.69, 4.92, 6.15 and 7.38 ms, a low excitation angle of 4° to minimize T1 bias, acceleration by CAIPIRINHA, factor 2 in both, phase-encoding and slice encoding directions, bandwidth 1080 Hz/pixel and acquisition time was 16 s in breath-hold (expiration). Image reconstruction for the calculation of PDFF maps was performed inline on the console of the scanner using a multistep fitting algorithm as described,7, 8 correcting for microscopic magnetic field inhomogeneities by correction for R2* of water and fat and the spectral components of fat.9 In this PDFF map, signal intensity values directly reflect PDFF of the respective tissue in percentages. Reproducibility and sensitivity of pancreatic fat measurements were recently reported,10 and showed that the applied technique is highly reproducible with coefficients of variation of <1% in pancreatic subregions (head, body, tail), and capable of detecting changes in the low range of PDFF. Following an overnight fast, participants were also subjected to 5-point 75 g OGTT to assess insulin sensitivity (Matsuda Index) and insulin secretion before and after treatment (Figure 1B). Insulin secretion was assessed by the oral Disposition Index as well as the areas under the C-peptide curves divided by the areas under the glucose curves during the first 30 min of the OGTT (AUC C-peptide0–30/AUC glucose0–30) as described previously.11 Statistical analyses were performed in JMP14. Changes within groups were compared by two-tailed paired t-tests, and treatment × time interactions were tested by repeated measures ANOVA. Values of p < .05 were considered statistically significant. Mean pancreatic fat content did not change in either treatment group [both p ≥ .2; before treatment: empagliflozin group 7.1 ± 4.6%, placebo group 10.3 ± 9.1; after treatment: empagliflozin group 6.2 ± 4.2%, placebo group 10.6 ± 8.7%; mean ± SD (Figure 2A)], and the course of pancreatic fat content was comparable between treatments (p = .2, Figure 2A,C). Insulin secretion, as assessed by Disposition Index (before treatment: empagliflozin group 731 ± 455 AU, placebo group 837 ± 596 AU; after treatment: empagliflozin group 765 ± 366 AU, placebo group 1073 ± 1040 AU, Figure 2B,D) and AUC C-peptide0-30/AUC glucose0–30 (adjusted for insulin sensitivity; before treatment: empagliflozin group 162 ± 50 AU, placebo group 166 ± 56 AU; after treatment: empagliflozin group 168 ± 35 AU, placebo group: 179 ± 65 AU) did not differ between treatments (P = 0.3 and P = 0.7, respectively). While empagliflozin treatment lowered fasting glucose significantly6 (Figure 2G), neither C-peptide nor glucose curves differed significantly after treatment (Figure 2E,F,H). In contrast to dietary energy restriction or exercise,1, 4, 12 empagliflozin did not reduce pancreatic fat in overweight and obese persons with prediabetes in our study. This implies that the beneficial effects of empagliflozin on systemic metabolism and general health5 do not depend on alterations in pancreatic fat. While we detected a reduction in liver fat and fasting glucose,6 our 8-week treatment with empagliflozin might have been not long enough to detect alterations in pancreatic fat, as we did not observe any changes in body weight either.6 However, glucagon-like peptide-1 receptor agonist treatment for 6 months, with significant but mild weight loss, was not able to reduce pancreatic fat content in patients with type 2 diabetes either.1 Improving β-cell function is thought to be pivotal for the prevention and remission of type 2 diabetes. The DiRECT lifestyle intervention trial detected a recovery of β-cell function as the key determinant of type 2 diabetes remission.4 Of note, β-cell recovery was dependent on pancreatic fat reduction.4 As empagliflozin was unable to reduce pancreatic fat significantly in our trial, it comes as no surprise that β-cell function did not improve in our participants with prediabetes. In overt diabetes, β-cell function is also compromised by glucotoxicity, which is typically not present in prediabetes. Via their glucose-lowering effect, SGLT2 inhibitors reduce glucotoxicity and can thereby improve insulin secretion in patients with overt diabetes.5 Although the glycaemic benefits of SGLT2 inhibitors have already been well established in diabetes, data from large randomized trials suggest that these benefits are inconsistent in prediabetes.13, 14 The uneven distribution of sexes among treatment groups posed a restriction to our work. Furthermore, the fact that the duration of the study was limited to 8 weeks meant that possible long-term effects of empagliflozin on pancreatic fat may have been overlooked. However, the effects on liver fat content and fasting glucose were detectable in our trial,6 which was sufficiently powered to detect the reduction of pancreatic fat as previously reported from the DiRECT trial4 (G*Power, α = 0.05, Power = 0.9). As the glucose-lowering effects of SGLT2 inhibitors are much stronger in patients with type 2 diabetes than what we have detected in prediabetes, stronger metabolic improvements could induce additional effects on pancreatic fat. However, Gaborit et al. did not detect effects of empagliflozin on pancreatic fat in this patient group.15 Insulin secretion (and insulin sensitivity) was not assessed via gold standard glucose clamps but estimated from OGTTs in our trial. Taken together, this randomized, controlled trial did not detect any effects of empagliflozin on pancreatic fat content and insulin secretion in persons with prediabetes. Further options for pancreatic fat reduction therefore need to be evaluated. JH, JM and CD researched and analysed data, SK, AP and RW researched data, ALB contributed to the discussion, HUH and AF contributed to the design of the trial and discussed data. MH researched data and supervised the project. All authors contributed to discussion and approved the final version of the manuscript prior to submission. We thank all patients for their participation in this trial. We particularly thank Dr Vera Valenta, Elisabeth Schrempf, Karin Waneck, Sabine Kümmerle, Ines Wagener, Eva-Maria Stehle and Alexandra Eberle (all Eberhard Karls Universität Tübingen) for their excellent technical assistance. Outside of the current work, SK reports lecture fees from Novo Nordisk. Outside of the current work, ALB reports lecture fees paid to the University of Tübingen from AstraZeneca, Boehringer Ingelheim and Novo Nordisk. He participated in advisory boards of Boehringer Ingelheim, Astra Zeneca and Novo Nordisk. Fees were paid to the University of Tübingen. Outside of the current work, AF reports lecture fees from Sanofi, Merck Sharp & Dohme and AstraZeneca. He participated in advisory boards of Boehringer Ingelheim, Sanofi, Novo Nordisk and Lilly. MH reports an independent research grant from Boehringer Ingelheim to the University Hospital Tübingen for this study. Outside of the current work, MH reports a research grant from Sanofi (to the University Hospital Tübingen), advisory board for Boehringer Ingelheim, lecture fees from Boehringer Ingelheim, Novo Nordisk and Amryt. All other authors declare that they have no competing interests. Professor Martin Heni is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. This study was pre-registered at Clinicaltrials.gov as trial number NCT03227484 and at The European Union Clinical Trials Register as EudraCT number 2016-003477-18. The peer review history for this article is available at https://publons.com/publon/10.1111/dom.14733. The data generated during the current study are not publicly available because they contain information that could compromise research participant privacy/consent.

Topics & Concepts