If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Exercise has opposite effects on heart FFA use in subjects with and without MAFLD.
•
Hepatic FFA utilization increases similarly after exercise.
•
VLDL-TG secretion and oxidation rates are greater in subjects with MAFLD.
•
Lactate increase during exercise is significantly impaired in MAFLD.
Abstract
Background and aims
Metabolic dysfunction-associated fatty liver disease (MAFLD) is associated with dyslipidemia and may promote cardiac lipotoxicity. Myocardial free fatty acids (FFA) oxidation (MOFFA) is normal in pre-diabetes, but reduced in heart failure. We hypothesized that during exercise MOFFA, very low-density lipoprotein triglycerides (VLDL-TG) secretion, hepatic FFA utilization, and lactate production differ among obese subjects with and without MAFLD.
Methods
Nine obese subjects with MAFLD and 8 matched subjects without MAFLD (Control) without a history of heart failure and cardiovascular disease were compared before and after 90-min exercise at 50% Peak oxygen consumption. Basal and exercise induced cardiac and hepatic FFA oxidation, uptake and re-esterification and VLDL-TG secretion were measured using [11C]palmitate positron-emission tomography and [1–14C]VLDL-TG.
Results
In the heart, increased MOFFA was observed after exercise in MAFLD, whereas MOFFA decreased in Control (basal vs exercise, MAFLD: 4.1 (0.8) vs 4.8 (0.8) μmol·100 ml−1 min−1; Control: 4.9 (1.8) vs 4.0 (1.1); μmol·100 ml−1 min−1, mean (SD), p < 0.048). Hepatic FFA fluxes were significantly lower in MAFLD than Control and increased ≈ two-fold in both groups. VLDL-TG secretion was 50% greater in MAFLD at rest and similarly suppressed during exercise. Plasma lactate increased significantly less in MAFLD than Control during exercise.
Conclusions
Using robust tracer-techniques we found that obese subjects with MAFLD do not downregulate MOFFA during exercise compared to Control, possibly due to diminished lactate supply. Hepatic FFA fluxes are significantly lower in MAFLD than Control, but increase similarly with exercise. VLDL-TG export remains greater in MAFLD compared to Control. Basal and post-exercise myocardial and hepatic FFA, VLDL-TG and lactate metabolism is abnormal in subjects with MAFLD compared to Control.
] and strongly associated with insulin resistant conditions, including abnormal free fatty acid (FFA) and lipoprotein metabolism, obesity, type 2 diabetes and increased atherosclerotic cardiovascular disease (ASCVD) [
]. In addition, since the flexibility in fuel selection between glucose and FFA is impaired in insulin resistant individuals, a disproportional increase in lipid oxidation arises [
]. The potentially harmful effects on heart metabolism might be further aggravated during exercise, where FFA availability is massively increased. Moreover, other fuel sources, such as lactate and ketone bodies, become increasingly important [
] and could unmask direct abnormalities in heart FFA oxidation as well as minute diastolic dysfunction in subjects with MAFLD without previous clinical experience of heart failure. Of note, improving insulin resistance, including FFA availability, is among the mechanisms by which physical exercise is proposed to improve MAFLD with and without cardiac lipotoxicity [
In subjects with MAFLD, hepatic FA oxidation, de-novo lipogenesis and very low-density lipoprotein triglycerides (VLDL-TG) secretion pathways are up-regulated which, if unbalanced, contribute directly to increased liver fat and greater circulating FFA and VLDL-TG levels [
]. Understanding how exercise affects myocardial and liver FFA metabolism in MAFLD may provide critical clues to early metabolic abnormalities that can predict development of heart failure.
We hypothesized, that aerobic exercise results in greater myocardial FFA oxidation (MOFFA) and uptake (MUFFA) rates in subjects with MAFLD compared to those without (Control). Additionally, subjects with MAFLD have decreased suppression of VLDL-TG secretion, increased VLDL-TG oxidation, and greater hepatic FFA uptake (HUFFA), oxidation (HOFFA) and esterification (HEFFA) rates during exercise than Control.
2. Patients and methods
2.1 Subjects
Informed consent from all subjects and Ethics Committee approval was obtained before the study, which was conducted in accordance with the 1975 Helsinki Declaration and registered at www.clinicaltrials.gov (NCT03583437). We identified MAFLD in accordance with the new definition by the presence of steatosis on magnetic resonance spectroscopy (MRS) and overweight/obesity [
]. We recruited 9 obese (BMI>28 kg/m2) subjects with MAFLD (fat fraction (FF%) >5.6% assessed from MRS and 8 matched subjects without MAFLD (Control). Additional inclusion criteria were age 40–70 years, non-smokers, no alcohol abuse (<21 units/week for men; <14 units/week for women), no history of cardiac or liver disease, clinical heart failure, ASCVD, type 2 diabetes, hyperlipidemia or chronic medical diseases, except hypertension (1 MAFLD, 3 Control). None received lipid lowering drugs. Four subjects took fish oil, which was paused 3 weeks before the study.
2.2 Study design
This was an open label interventional trial with parallel design comparing subjects with MAFLD and metabolically healthy obese Control subjects during rest and acute exercise. Eligible subjects were invited for a screening visit. A magnetic resonance imaging (MRI) and MRS to assess abdominal fat distribution and liver fat content, and a Fibroscan® to measure liver stiffness were obtained. All had a normal blood and chemistry panel and ECG documented before inclusion.
One week before the study a 12-h overnight fasting blood sample was obtained under aseptic conditions for VLDL-TG tracer preparation and a dual-X-ray absorptiometry (DEXA) scan was performed. Peak consumption uptake ( O2peak) was determined by an incremental (20 W/min) cycling protocol until exhaustion. Finally, a dietitian interviewed each subject and designed weight maintaining diets (55% carbohydrate, 15% protein, and 30% fat), which was provided by the hospital kitchen the 3 days preceding the study day [
]. All were instructed to avoid strenuous physical activity during this period.
2.3 Metabolic study day
Subjects were admitted to the Clinical Research Unit (CRU) at 10 p.m. the evening before the study. Only tap water was allowed and they remained in bed under thermoneutral conditions. The study day (Fig. 1) comprised a 4-h basal period (0–240 min), including a dynamic [11C]palmitate positron-emission tomography (PET) computed tomography (CT) scan imaging (155–205 min), followed by a 1½-hour cycling exercise period at 50% of O2peak (240–330 min), immediately followed by a second [11C]palmitate PET/CT scan (345–395 min) and a 3½-hour recovery period (330–540 min). Catheters were placed in an antecubital vein for infusions and a contralateral heated hand vein for arterialized blood in the first 4 subjects. Due to difficulties with keeping iv catheters in place, the subsequent 13 subjects had an arterial cannula placed in a wrist artery. A primed-constant infusion of [1–14C]triolein labeled VLDL-TG (20% bolus, 80% constant) was administered during the whole study period to measure VLDL-TG kinetics. Blood samples for VLDL-TG concentration and specific activity (SA) were collected together with 14CO2 in breath samples at 10 min intervals during the last 30 min of each steady-state period. A 1-h constant infusion of [9,10-3H]palmitate was given from t = 170–230 min and t = 270–330 min to measure palmitate turnover. Blood samples for palmitate concentration and SA were collected before infusion and every 10 min during the last 30 min of each infusion period. Indirect calorimetry was performed from t = 90–105 min, t = 270–285 min and t = 460–475 min. Insulin and metabolite concentrations were collected at frequent intervals. After the study all catheters were removed, and the subjects discharged.
]. LCModel software package version 6.3-1L (Stephen Provencher, 2016) was used to analyze data. The cut-off between normal and abnormal intrahepatic TG (IHTG) content was determined from the Dallas Heart study [
] to be FF% >5.6%. Lean liver mass was calculated as liver volume (ml) x (1-FF%).
2.5 Fibroscan®
Fibroscan® was used to measure liver stiffness by transducer probe-induced elastic share wave that propagates through liver tissue expressed in kPa.
2.6 Body composition
Total body fat, fat percentage and lean body mass (LBM) were determined by DEXA scanning (QDR-2000, Hologic).
2.7 [11C]palmitate PET/CT scan
PET/CT scans were conducted to measure myocardial and hepatic FFA kinetics using either a Siemens Biograph TruePoint 64 (first three subjects) or a Siemens Biograph Vision 600 PET/CT (Siemens). In all patients, a low-dose CT scan was obtained for attenuation and anatomic localization purposes.
2.7.1 PET protocols and data acquisition
The resting [11C]palmitate scan was initiated at t = 155 min as a bolus (410 MBq ±3 MBq) and a 50 min list mode scan (frame structure 6x5, 6x10, 3x20, 5x30, 5x60, 8x150 and 4 × 300 s) was performed. The post-exercise PET/CT scan was initiated at t = 345 min as a bolus (417 MBq ±5 MBq) with same frame structure as the resting scan. For the Truepoint 64 system an Ordered-subset-expectation-peakimation 3D reconstruction with 3 iterations, 21 subsets, 5 mm Gaussian post-filter, 168x168 matrix was used, whereas for the Vision system we used Point-of-spread + Time-of-flight reconstruction with 4 iterations, 5 subsets, 128x128 matrix and a 5 mm Gaussian postfilter. All dynamic PET data were decay corrected to scan start.
2.7.2 PET image analysis
2.7.2.1 Myocardial FFA metabolism
Myocardial FFA metabolism was analyzed using aQuant cardiac software (MedTrace, Denmark) and a three-tissue compartment model in which three rate constants need to be fitted. The macro-parameters MOFFA, myocardial FFA esterification (MEFFA) and MUFFA were defined according to Bergmann et al. [
For resting myocardial palmitate kinetics, the input function was corrected for [11C]-metabolites using population-based estimates obtained in our own facility [
]. Since no published data exist for post-exercise palmitate metabolite fractions, these were performed individually using the same method as described above [
For hepatic palmitate kinetics, the input function (Ca) was calculated based on dual input from hepatic artery (aorta volume-of-interest (VOI)) and portal activity (portal VOI drawn on summed images from 30 to 120 s) using PMOD 3.8 software (Zurich, Switzerland). Hepatic parenchymal VOI was drawn on 8 consecutive slices in the right liver lobe avoiding large vessels and bile drainage. Image derived activity was corrected for hematocrit to obtain plasma activity estimates. Hepatic arterial supply was assumed to constitute 20% of total liver blood supply. The dual-input function was corrected for metabolites as described above and elsewhere [
]. In short, the model consists of 3 tissue compartments, representing free [11C]palmitate, [11C]palmitate bound in complex lipids, and [11C]-oxidative breakdown products. The exchange of radioactivity across compartments is described by 4 rate-constant terms. The differential equations were solved analytically by in-house developed software (ifit) and the rate constants were then used to calculate hepatic FFA fluxes (μmol⋅ml tissue−1⋅min−1):
Where k = rate-constant, p = plasma.
2.8 VLDL-TG kinetics
Basal and exercise steady-state VLDL-TG kinetics (secretion and oxidation) were calculated as previously described [
]. An 80 ml blood sample was obtained and VLDL-TG labeling for autologous infusion was performed using 20 μCi of [1–14C]triolein (PerkinElmer) as described previously [
A 1-h constant infusion of [9,10-3H]palmitate (0.3 μCi/min) was employed to measure systemic palmitate turnover. Plasma palmitate concentration and SA were measured by HPLC using [2H31] palmitate as internal standard [
]. Steady-state SA was verified for each individual. Palmitate turnover (μmol/min) was calculated as [9,10-3H]palmitate infusion rate (dpm/min) divided by the steady-state palmitate SA (dpm/μmol).
2.10 Echocardiography
A 2D and Doppler examination with a GE vivid E95 echocardiography was performed to measure left ventricle (LV) end-diastolic volume, end-systolic volume, mass, and mass-index. Relative wall thickness was calculated as (2 x posterior wall thickness at end diastole)/LV diastolic dimension. The Simpson method was used to calculate LV ejection fraction. Cardiac output was defined as the time-velocity integral of the LV outflow x LV outflow tract area. Diastolic function was calculated by measurement of trans-mitral flow and tissue Doppler parameters including the early (E) and late (A) diastolic filling velocities, the E/A ratio and E’ of the lateral part of the annular disc.
2.11 Indirect calorimetry
Indirect calorimetry (Oxycon Pro, Erich Jaeger (16 subjects) and Deltatrac monitor, Datex Instrumentarum (1 subject)) was used to measure energy expenditure (EE) and substrate oxidation rates [
]. Urine was collected during the basal period to calculate protein oxidation.
2.12 Laboratory procedures
An YSI 2.300 STAT Plus glucose analyzer (YSI) was used to analyze plasma glucose and lactate. The Homeostatic Model Assessment for insulin resistance (HOMA-IR) was calculated using the standard formula. Triglycerides were measured on a Cobas 111 using glycerol blanked kits, serum insulin with an immunoassay (DAKO Denmark A/S), and serum FFA by a colorimetric method. Plasma Pro-Brain Natriuretic Peptide (Pro-BNP) was analyzed using standard laboratory testing and plasma Beta-hydroxybutyrate (BOH) using hydrophilic interaction liquid chromatography electrospray tandem mass spectrometry [
Simultaneous determination of β-hydroxybutyrate and β-hydroxy-β-methylbutyrate in human whole blood using hydrophilic interaction liquid chromatography electrospray tandem mass spectrometry.
Metformin does not affect postabsorptive hepatic free fatty acid uptake, oxidation or resecretion in humans: a 3-month placebo-controlled clinical trial in patients with type 2 diabetes and healthy controls.
], whereas post-exercise PET-FFA kinetics is available. We calculated a sample of n = 6 in each group to detect a mean (SD) difference of 1.0 (0.5) μmol·100 ml tissue−1·min−1 (α = 0.05, β = 0.80) in baseline PET-FFA oxidation. Moreover, to detect a difference in the percent post-exercise change in PET-FFA oxidation of 20 (10)% (α = 0.05, β = 0.80) we also calculated n = 6 in each group. Since we previously found that n = 8 is needed to detect similar differences in VLDL-TG kinetics [
] we chose to include 8 subjects in each group. Normality was evaluated by QQ-plots and Shapiro-Wilk normality test. During steady-state periods, kinetics are analyzed as average values. Between- and within-group comparisons and correlations were analyzed using paired/unpaired t-test or Mann-Whitney/Wilcoxon's test as appropriate. Comparisons between groups and study periods were performed using a linear mixed model with maximum likelihood estimation and time and group as factors and time vs group as interaction term. p values < 0.05 were considered significant.
3. Results
3.1 Subjects
Table 1 shows subjects characteristics. Subjects with MAFLD had significantly greater HbA1c and HOMA-IR. The O2 during exercise was similar in the two groups, averaging 50% of O2peak, and the workload was also similar (MAFLD: 61 (12) Watt; Control: 76 (23) Watt, p = 0.11). One MAFLD only completed the basal PET-FFA and palmitate turnover measurements due to nausea during exercise. In the same subject and in 1 Control the VLDL-TG kinetic were not obtained for technical reasons. One Control was moving during the PET/CT scan after exercise, resulting in invalid data for liver PET/CT.
Data are mean (SD) or median (range). *p < 0.01; ALAT: alanine aminotransferase; VAT: visceral adipose tissue; SAT: abdominal subcutaneous adipose tissue.
Glucose, insulin, lactate, BOH, total-TG and VLDL-TG concentrations are presented in Fig. 2. Subjects with MAFLD had significantly greater glucose, insulin, total-TG and VLDL-TG during the basal period compared with Control. No significant interactions with exercise were found between groups with respect to glucose, insulin, BOH, TG and VLDL-TG. Lactate concentrations increased significantly in both groups during exercise. The increase in lactate was significantly less in MAFLD than Control. Basal BOH was greater, but not significantly, in the Control than MAFLD group and decreased slightly to similar levels during exercise (Fig. 2D).
Basal MOFFA and MUFFA were slightly, but not significantly, lower in MAFLD compared with Control, while basal MEFFA was similar in the two groups (Fig. 3A–C; exact numbers given in Supplementary Table 1). Immediately after exercise MOFFA changed significantly in opposite directions in the two groups with an increase in MAFLD and a decrease in Control. Similar changes were noted in MUFFA. MEFFA increased similarly and significantly in both groups.
Oxidation (A), uptake (B) and esterification (C) rates in the basal state and after 90 min exercise. Open symbols: MAFLD, filled symbols: Control. *p < 0.05 (interaction), †p < 0.005 (time effect).
Basal HOFFA and HUFFA were similar in the two groups, whereas basal HEFFA was significantly lower in MAFLD (p < 0.05) (Fig. 4A–C; Supplementary Table 1). The HOFFA, HUFFA and HEFFA increased similarly and significantly in the two groups following exercise. HOFFA tended to increase less in MAFLD than Control, but the interaction with exercise was not significant. The levels of both HOFFA, HUFFA and HEFFA were significantly lower in MAFLD than Control.
Oxidation (A), uptake (B) and esterification (C) rates in the basal state and after 90 min exercise. Open symbols: MAFLD, filled symbols: Control. *p < 0.05 (time effect).
Basal VLDL-TG secretion was greater in MAFLD compared with Control (p < 0.01) and was similarly suppressed during exercise and recovery (Fig. 5A). Normalizing VLDL-TG secretion to LBM eliminated the group effect. VLDL-TG oxidation rate was greater in MAFLD compared with Control in the basal state (p < 0.01) and during the whole study day (Fig. 5B). The oxidation rate increased similarly during exercise and decreased to pre-exercise levels during recovery. Again, normalizing to LBM eliminated the group effect.
Basal palmitate flux was similar between groups and increased similarly during exercise (MAFLD: 340 (86) vs. 566 (158) μmol·min−1; Control: 263 (56) vs. 576 (168) μmol·min−1, basal vs exercise, respectively, p < 0.001 time effect, mixed model). Palmitate SA was stable during the last 30 min of exercise indicating a constant palmitate clearance rate. No significant correlations were observed between palmitate flux and myocardial or hepatic PET-FFA fluxes.
3.7 Whole body glucose and lipid oxidation
We found significant and similar increaments in glucose and lipid oxidation rates during exercise with a return towards pre-exercise levels during recovery in the two groups. Exact numbers are shown in Supplementary Table 2.
3.8 Echocardiography
We found similar results in the two groups, except for a significantly lower left atrium area (LA area) and E/A ratio in the group with MAFLD. Exact values are presented in Supplementary Table 3.
4. Discussion
The present study is the first to simultaneously assess myocardial and hepatic FFA utilization and VLDL-TG secretion during rest and moderate-intensity aerobic exercise in subjects with and without MAFLD and no history of clinical heart failure or ASCVD. The study used a novel approach by combining robust and validated tracer techniques, [11C]palmitate PET/CT and triolein labeled VLDL-TG. The main and novel finding is that exercise induces opposite effects on myocardial FFA utilization, with an increase in FFA oxidation and uptake in MAFLD as opposed to a decrease Control. Moreover, hepatic FFA utilization increases after exercise, but are overall diminished in MAFLD compared to Control. Systemic palmitate turnover is not different between groups and does not correlate with PET-FFA kinetics. In addition, hepatic VLDL-TG secretion and oxidation rates are greater in MAFLD than Control, but suppressed similarly during exercise in both groups. Finally, the study demonstrates that the increase in plasma lactate during exercise is significantly impaired in MAFLD compared to Control.
The oxidation of FFA in the heart was strikingly different between the two groups. While MOFFA decreased immediately after exercise in Control, an increase was observed in MAFLD. This may relate to the blunted increase in lactate in MAFLD. During rest, lactate accounts for <10% of myocardial energy consumption, while FFA constitute the major energy source [
]. It is well recognized that plasma lactate increases during moderate-intensity exercise, resulting in increased lactate uptake and oxidation in the heart of healthy subjects [
]. Therefore, impaired lactate production during exercise may lead to increased dependence of the myocardium on other substrates, such as FFA. The increase in skeletal muscle and adipose tissue lactate release during insulin stimulation of glycolytic flux is impaired in insulin-resistant, obese women [
]. Hence, maybe because of greater insulin sensitivity, lactate appears to contribute more as a substrate for myocardial oxidation in the Control group compared with MAFLD and may even serve to protect the myocardium against accumulation of toxic lipid intermediates from incomplete FFA oxidation, which can promote insulin resistance, oxidative stress and inflammation [
]. Thus, our study proposes potential mechanistic evidence for a risk of FFA induced development of cardiomyopathy secondary to increased uptake and oxidation of FFA during and after physical activity. If such mechanisms operate in humans with MAFLD, this represents a metabolic paradox, since exercise is generally recommended for improvement of insulin resistance and cardiometabolic health.
The subjects in this study had no history of clinical heart failure or ASCVD. Even though echocardiographic measurements were almost similar in the two groups, there was significantly increased LA area and reduced MV E/A ratio in MAFLD compared to Control, both hallmarks of diastolic dysfunction [
Altered myocardial substrate metabolism and decreased diastolic function in nonischemic human diabetic cardiomyopathy: studies with cardiac positron emission tomography and magnetic resonance imaging.
]. In subjects with clinical heart failure, however, reduced cardiac oxidation of FFA, glucose, lactate and branched amino acids, and increased oxidation from ketone bodies and glycolysis has been reported [
]. Of note, basal MOFFA and MUFFA rates in our groups were similar to those reported in subjects with impaired glucose tolerance and their respective control subjects [
Impaired free fatty acid uptake in skeletal muscle but not in myocardium in patients with impaired glucose tolerance: studies with PET and 14(R,S)-[18F]fluoro-6-thia-heptadecanoic acid.
]. Therefore, we speculate that the novel observation of an opposite effect on MOFFA, i.e. an increase in MAFLD and a decrease in Control, in subjects without clinical heart failure and ASCVD may represent an early sign of future heart failure among subjects with MAFLD.
In contrast, hepatic FFA utilization was significantly reduced in MAFLD compared with Control, with a similar increase immediately after exercise, most likely due to exercise induced lipolysis and hepatic FFA delivery. Elevated FFA levels can cause defects in the hepatic oxidation in insulin resistant, obese subjects, implying hepatocyte injury via the production of toxic metabolites and oxidative stress [
], which may promote progression to NASH. Of note, palmitate flux was not different between the two groups, but we cannot exclude that hepatic FFA delivery differed between the MAFLD and Control subjects. However, exercise-induced suppression of splanchnic bloodflow is similar at comparable exercise intensity relative to O2peak [
], as in the present study. Finally, we found significantly greater VLDL-TG secretion and oxidation rates during rest and exercise in MAFLD compared with Control, with a similar suppression of VLDL-TG secretion during exercise and recovery, which extends our previous findings in subjects with MAFLD [
] and supports a role for splanchnic bloodflow reduction during exercise and early recovery. Collectively, our results suggest, that in the presence of MAFLD the liver upregulates yet unresolved mechanisms involved in minimizing excess intrahepatic lipid accumulation by reducing FFA uptake and increasing VLDL-TG export.
A significant strength of this study is the comprehensive in vivo characterization of the effects of exercise on myocardial and hepatic FFA and VLDL-TG metabolism by using robust methods with validated kinetic models. Moreover, O2peak was similar in the two groups and therefore the workload (50% of O2peak) on the study day was also similar. In addition, the subjects were provided isocaloric meals for three days and were admitted to the CRC the evening before the study in order to minimize day-to-day variability of palmitate and VLDL-TG kinetics [
]. The study also has limitations. First, the sample size is small, so type 2 errors cannot be excluded. Second, MAFLD was determined by MRS, not biopsies. Third, we did not measure heart lactate or glucose oxidation directly, but we propose this be tested using a lactate PET-tracer. In addition, glucose oxidation accounts for a rather small percentage of cardiac oxidation [
]. Fourth, subjects with MAFLD were on average 5 years older than Control. We cannot determine whether the subjects with MAFLD had only simple steatosis when they were 5 years younger. Since the O2peak and workload did not differ between groups we do not believe that the age difference implicated major differences in substrate oxidation. Fifth, as sex-differences in FFA and VLDL-TG metabolism are well established [
] our results might be different if the study was restricted to either men or women, which deserves further investigations.
In conclusion, these novel findings indicate unresolved mechanisms whereby subjects with MAFLD are able to minimize exercise induced liver lipid overload, yet fail to protect the heart against increased FFA uptake and oxidation, maybe due to reduced lactate availability. Further large scale intervention studies are needed to examine the pathophysiological implications for the heart.
Financial support
This study was supported by The Novo Nordic Foundation (grant numbers NNF18OC0031804, NNF16OC0021406), The Independent Research Fund Denmark (grant number 8020-00420B), Aarhus University and Steno Diabetes Centre Aarhus, Denmark. Funding sources had no involvement in any part of this study. All authors have read and approved the final manuscript.
Data statement
The datasets generated during and/or analyzed during the current study are available from the corresponding author upon reasonable request.
CRediT authorship contribution statement
Jeyanthini Risikesan: Methodology, Validation, Formal analysis, Investigation, Resources, Data curation, Writing – original draft, Writing – review & editing, Visualization. Sara Heebøll: Investigation, Writing – review & editing. Indumathi Kumarathas: Investigation, Writing – review & editing. Kristian L. Funck: Methodology, Formal analysis, Investigation, Resources, Data curation, Writing – review & editing. Esben Søndergaard: Investigation, Writing – review & editing. Rakel F. Johansen: Investigation, Writing – review & editing. Steffen Ringgaard: Methodology, Formal analysis, Investigation, Resources, Data curation, Writing – review & editing. Lars P. Tolbod: Methodology, Software, Validation, Formal analysis, Resources, Data curation, Writing – review & editing. Mogens Johannsen: Methodology, Formal analysis, Investigation, Resources, Data curation, Writing – review & editing. Helle L. Kanstrup: Methodology, Formal analysis, Investigation, Resources, Data curation, Writing – review & editing. Henning Grønbæk: Methodology, Resources, Writing – review & editing, Supervision. Jan Frystyk: Writing – review & editing, Supervision, Funding acquisition. Lars C. Gormsen: Methodology, Software, Validation, Formal analysis, Investigation, Resources, Data curation, Writing – original draft, Writing – review & editing, Visualization, Supervision. Søren Nielsen: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Resources, Data curation, Writing – original draft, Writing – review & editing, Visualization, Supervision, Project administration, Funding acquisition.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
The authors thank the research volunteers. We also thank Susanne Sørensen, Lone Kvist and Eva Schriver for technical assistance.
Appendix A. Supplementary data
The following is the Supplementary data to this article.
Simultaneous determination of β-hydroxybutyrate and β-hydroxy-β-methylbutyrate in human whole blood using hydrophilic interaction liquid chromatography electrospray tandem mass spectrometry.
Metformin does not affect postabsorptive hepatic free fatty acid uptake, oxidation or resecretion in humans: a 3-month placebo-controlled clinical trial in patients with type 2 diabetes and healthy controls.
Altered myocardial substrate metabolism and decreased diastolic function in nonischemic human diabetic cardiomyopathy: studies with cardiac positron emission tomography and magnetic resonance imaging.
Impaired free fatty acid uptake in skeletal muscle but not in myocardium in patients with impaired glucose tolerance: studies with PET and 14(R,S)-[18F]fluoro-6-thia-heptadecanoic acid.
00117-X&id=ga1.jpg){kind=link}
00117-X&id=gr1.jpg){kind=link}
00117-X&id=gr2.jpg){kind=link}
00117-X&id=gr3.jpg){kind=link}
00117-X&id=gr4.jpg){kind=link}
00117-X&id=gr5.jpg){kind=link}