Bariatric Surgery Rapidly Decreases Cardiac Dietary Fatty Acid Partitioning and Hepatic Insulin Resistance Through Increased Intra-abdominal Adipose Tissue Storage and Reduced Spillover in Type 2 Diabetes

Bariatric surgery is the most effective treatment for type 2 diabetes (T2D) in severely obese individuals (1,2). Its impact on metabolism goes beyond the effects of weight loss as improvements in glucose metabolism have been shown as early as 3 days after surgery (3–6). Improved fasting glucose and postprandial glucose tolerance is seen early after all current bariatric procedures (6). Within the first few days to weeks, reductions in hepatic insulin resistance and endogenous glucose production (EGP) occur, whereas peripheral insulin resistance improves only later, after substantial weight loss has occurred (3–9).

Besides abnormal glucose metabolism, T2D is associated with impaired lipid metabolism and is characterized by excess fat deposition in lean organs (10). The excessive exposure of lean tissues to long-chain fatty acids creates a lipotoxic environment contributing to insulin resistance and pancreatic β-cell dysfunction (11,12). Obesity and the metabolic syndrome are associated with adipose tissue metabolic dysfunction, characterized by decreased storage capacity of dietary fatty acids (DFAs) and adipose tissue resistance to insulin-induced inhibition of intracellular triglyceride (TG) lipolysis (13). This adipose tissue metabolic dysfunction contributes to excess exposure to fatty acids and the development of insulin resistance in lean tissues (14).

We previously showed, using positron emission tomography coupled with computed tomography (PET/CT) and oral administration of 14-R,S-[¹⁸F]-fluoro-6-thia-heptadecanoic acid ([¹⁸F]-FTHA), that individuals with impaired glucose tolerance have decreased uptake of DFA in abdominal subcutaneous and intra-abdominal adipose tissues as well as increased cardiac DFA uptake associated with reduced left ventricular function (15). This suggests that impaired adipose tissue DFA storage could lead to impaired myocardial function in patients with prediabetes and T2D via increased myocardial exposure and uptake of circulating DFA.

The current study investigated the effect in T2D of bariatric surgery on DFA metabolism and organ partitioning within 2 weeks after the operation. We hypothesized that increased cardiac DFA uptake and partitioning associated with T2D would be reduced after bariatric surgery together with increased adipose tissue DFA uptake and decreased DFA spillover.

The study participants were part of two observational longitudinal trials on the mechanisms of T2D remission after bariatric surgery (NCT02815943 and/or NCT02390973– REMISSION trial). Participants with T2D undergoing sleeve gastrectomy (SG) or SG with biliopancreatic diversion with duodenal switch (BPD-DS) (see Marceau et al. [16] for a description of BPD-DS) at the Institut universitaire de cardiologie et de pneumologie de Québec (IUCPQ) were recruited through the preoperative bariatric surgery clinic. Patient selection for bariatric surgery followed standard National Institutes of Health consensus recommendations (i.e., patients aged 18–65 with a BMI >40 kg/m² or BMI >35 kg/m² with comorbidities. Exclusion criteria included 1) use of thiazolidinedione, insulin, or fibrates; 2) contraindication to stop statins, antidiabetic, or antihypertensive medications before the metabolic study visits; 3) past medical history of significant cardiac, liver (other than fatty liver disease or nonalcoholic fatty liver disease without cirrhosis), renal, or other systemic diseases; 4) early surgical complication that contraindicated participation in the second postoperative study; 5) exposure to other imaging that included ionizing radiations during the previous year (only for the PET/CT procedures). T2D was diagnosed based on Diabetes Canada guidelines (https://guidelines.diabetes.ca/cpg/chapter3). Informed written consent was obtained from all participants in accordance with the Declaration of Helsinki, and the protocol received approval from the Human Ethics Committee of the Centre de recherche de l'IUCPQ and the Centre de recherche du CHUS.

Participants underwent two identical metabolic studies: 1) within 1 month before bariatric surgery (preoperative) and 2) between 8 and 12 days after bariatric surgery (12D postoperative). Participants followed 3 days of an isocaloric diet before the preoperative test, and the standard diet prescribed after bariatric surgery was followed before the postoperative test. On both occasions, after a 15-h fast, a standard liquid meal (80 mL) containing 50% of energy (43 g) from glucose, 30% (9.4 g) from fat, and 20% (17.8 g) from proteins was administered in four aliquots of 20 mL over 20 min. The standard liquid meal was supplemented with 0.9 g of 6,6-[²H]glucose and 9 μmol/kg of [U-¹³C]palmitate (Cambridge Isotope Laboratories). A primed (1.5 μCi) continuous (0.5 μCi/min) intravenous perfusion of 6-[³H]glucose (17) was started 180 min before and continued up to 6 h after the liquid meal intake. A continuous intravenous perfusion (0.010 µmol/kg/min) of [7,7,8,8-²H]palmitate (Cambridge Isotope Laboratories) mixed in 25% albumin was started 60 min before and continued up to 6 h after meal intake (5). Blood was drawn in collection tubes (containing dipeptidyl-peptidase-4 inhibitors) at times −180, 0, 10, 20, 30, 40, 50, 60, 75, 90, 120, 150, 180, 210, 240, 300, and 360 min for quantification of tracer enrichment and metabolite or hormone levels. A bolus infusion of sterile 1-[¹³C]NaH[¹³C]O₃ (1.2 μmol/kg) (Cambridge Isotope Laboratories) to prime the bicarbonate pool was administered at time −60 min. Breath samples were collected at fasting and every hour postprandially.

Organ-specific uptake and partitioning of DFA was assessed using the oral [¹⁸F]-FTHA PET/CT method (15,18–21). [¹⁸F]-FTHA given orally is absorbed and metabolized similarly to long-chain DFAs, assembled in chylomicron (CM)-TG and, after intravascular lipolysis, can be taken up locally by tissues and recirculated as nonesterified fatty acids (NEFA) via the DFA spillover pathway from adipose tissues and as VLDL-TG after liver uptake of [¹⁸F]-FTHA (18). Oral [¹⁸F]-FTHA is absorbed and distributed into CM-TG at the same rate and to the same extent as oral [³H]triolein (18).

A gel capsule (T.U.B. Enterprises) containing ∼70 MBq of [¹⁸F]-FTHA (18) mixed in Intralipid 20% (Baxter, Mississauga, Ontario, Canada) was given orally at time 0 of the liquid meal in a subset of eight patients because of contraindication to the PET/CT imaging in the other participants (n = 4 with exposure to ionizing radiations in the year before the metabolic study). A dynamic list-mode PET acquisition centered on the thoracoabdominal segment (15 × 120 s) was performed between time 150 and 180 min, and a whole-body PET/CT acquisition was performed at time 360 min (15,18–20). The total radioactivity exposure to the participants was <20 mSv, including the two protocols. All tracers were tested for sterility and pyrogenicity.

For dynamic PET acquisitions, regions of interest were manually drawn on the liver, left ventricle, and aorta to generate tissue/blood time-radioactivity curves. Myocardial and hepatic fractional DFA uptakes (K_i) were determined using Patlak linearization (15,19). For the whole-body PET acquisition, regions of interest (left ventricle, liver, vastus lateralis, subcutaneous abdominal and upper thigh adipose tissue, and left perirenal adipose tissue) were manually drawn over the CT images, and the mean value of pixels (mean standard uptake value [SUV]) for all tissues of interest was recorded (15,18–20).

Insulin, C‐peptide, leptin, glucagon, and glucose-dependent insulinotropic peptide (GIP) were measured by Luminex LX200 multiplex immunoassays (Millipore, Billerica, MA), peptide YY (PYY) by radioimmunoassay (Millipore), and total glucagon-like peptide-1 (GLP‐1) by ELISA (Millipore) A colorimetric assay (Cayman Chemical, Ann Arbor, MI) was used to quantify β-hydroxybutyrate plasma levels. CM and VLDL were isolated by gradient-density ultracentrifugation, and total plasma and CM [¹⁸F] activities were measured (18). Plasma and meal [6,6-²H]glucose enrichment was measured by liquid chromatography–tandem mass spectrometry and [³H]glucose activity was determined by liquid scintillation (22) to calculate total, endogenous, and meal Ra_glucose (23). [U-¹³C]palmitate and [7,7,8,8-²H]palmitate enrichments were determined using gas chromatography–tandem mass spectrometry to calculate plasma palmitate and NEFA appearance rates (5,24). Breath [¹³C]O₂-to-[¹²C]O₂ ratio was determined by isotope ratio mass spectrometry (Sercon Ltd., Crewe, Cheshire, U.K.). The fraction of oxidized [U-¹³C]palmitate from the diet, NEFA spillover rate, and Ra_NEFA from adipose tissue intracellular lipolysis were calculated as described (20,24)

Briefly, total Ra_palmitate and Ra_NEFA were calculated by intravenous palmitate infusion rate × plasma tracer-to-tracer ratio (TTR) (total Ra_palmitate) and multiplied by the plasma NEFA/plasma palmitate concentration to obtain the total Ra_NEFA. Ra_NEFA spillover was calculated from total Ra_palmitate × plasma TTR [¹³C]palmitate/CM TTR [¹³C]palmitate × total fatty acids/palmitate concentration in CMs. Ra from adipose tissue lipolysis (nonspillover Ra_NEFA) was calculated as total Ra_NEFA − Ra_NEFA spillover.

Orally administered postprandial [¹⁸F]-FTHA and [¹³C]palmitate excursion in total plasma, CMs, and VLDL are reported from 60, 120, and 180 min, respectively, because of their small concentration levels at prior times, Ra_NEFA spillover and nonspillover Ra_NEFA (i.e., total Ra_NEFA − Ra_NEFA spillover) are reported from postprandial time 120 min because they are calculated from CM tracer enrichment.

The oral minimal model Si was calculated using glucose and insulin levels at times 0, 10, 20, 30, 60, 90, 120, 150, 180, 240, 300 and 360 using SAAM II version 2.3 software (The Epsilon Group). Hepatic insulin resistance was determined by the total Ra_glucose at fasting (EGP) multiplied by fasting insulin and adipose tissue insulin resistance by Ra_NEFA at fasting (endogenous Ra_NEFA) multiplied by fasting insulin. Body composition was assessed with a bioelectrical impedance scale in light clothes the morning before each metabolic study (Tanita Corporation of America, Arlington Heights, IL).

Indirect calorimetry (Vmax29n, Sensormedics and Promethion High-Definition Room Calorimetry System; Sable Systems Int.) was performed at fasting and every hour after the liquid meal test. Fasting and postprandial protein, carbohydrate, and lipid oxidation rates (in g/min) were calculated as described previously (25).

Data are expressed as median and interquartile range (IQR). Continuous variables were analyzed using a paired Wilcoxon test for the data before and after bariatric surgery. Two-way ANOVA for repeated measures or mixed model (unbalanced data) with effect of bariatric surgery, postprandial time, and interaction as the independent variables was used to analyze differences in plasma metabolites throughout the postprandial period, with the Sidak test to correct for multiple comparisons. A two-tailed P value <0.05 was considered statistically significant. All analyses were performed with JMP software for Windows, version 7.0 (SAS Institute, Cary, NC), or GraphPad Prism version 7.0 and 8.0 (GraphPad Software, San Diego, CA).

The data sets generated during and/or analyzed during the current study are available from the corresponding author upon reasonable request. No applicable resources were generated or analyzed during the current study.

The metabolic studies were performed in 12 participants (7 women, 5 men) with T2D, with a median age of 48 years (IQR 36–55), before and 8–12 days after surgery (9 SG, and 3 BPD-DS).

The median (IQR) weight and BMI before surgery was 121.8 kg (94.1–137.9) and 43.6 kg/m² (40.5–49.1), respectively, and 115.6 kg (89.9–130.3) and 41.2 kg/m² (38.4–46.4) after surgery, for a median weight loss of 5.6% of total body weight (Supplementary Table 1). Lean body mass was 59.3 kg (49.2–74.5) before surgery and 58.4 kg (48.7–62.3) after surgery (P = 0.04). There was no difference in the percentage of body fat mass before versus after surgery (48.7% [43.6–54.4] vs. 47.0% [44.5–51.78], P = 0.53).

After bariatric surgery, there was a significant decrease in cardiac DFA partitioning, with SUV 1.75 (1.39–2.57) before vs. 1.09 (1.04–1.53) after surgery (P = 0.01) (Fig. 1A). Before the surgery, five of the eight participants had cardiac DFA partitioning over the upper bound of the 95% CI of the mean of normal-weight healthy participants with normal glycemia who received the same oral dose of [¹⁸F]-FTHA (15). In all but one, cardiac DFA partitioning was lowered within or below the 95% CI of the mean of healthy participants within 12 days after surgery. There was no difference in DFA partitioning after bariatric surgery in the other lean tissues assessed (liver, P > 0.99; or muscle, P = 0.19) (Fig. 1B and C). Along with decreased myocardial DFA partitioning, there was a significant increase in perirenal adipose tissue, a surrogate for intra-abdominal fat depots, with SUV 0.15 (0.04–0.31) before vs. 0.49 (0.20–0.59) after surgery (P = 0.008) (Fig. 1D), reflecting increased postprandial storage of DFA per volume of tissue. We did not find differences in DFA partitioning in abdominal or lower body subcutaneous adipose tissues (P > 0.58) (Fig. 1E and F).

Interestingly, there was no difference in cardiac or liver uptake of DFA between 150 and 180 min postprandially from dynamic PET acquisition at a time when DFAs are transported in the circulation mainly as CM-TG (Fig. 1G and H). Furthermore, there was a slight increase in total plasma [¹⁸F]-FTHA activity (interaction P < 0.05) and no change in CM [¹⁸F]-FTHA activity, in recovery of [U¹³C]palmitate from the meal in CMs or VLDLs, or in CM-TG and VLDL-TG levels throughout the postprandial period after bariatric surgery (Fig. 2).

After bariatric surgery, there was a decrease in the postprandial inhibition of the plasma NEFA concentration and in the total Ra and nonspillover Ra_NEFA (Fig. 3A–C). However, NEFA spillover from DFA was very significantly reduced after surgery (Fig. 3D). The total area under the curve (AUC) of DFA spillover decreased from 56.7 before to 24.7 mmol over 6 h after meal intake after surgery (P = 0.01).

There was a slight reduction in resting energy expenditure after bariatric surgery (Fig. 4A) and a significant decrease in the fasting protein oxidation rate (Fig. 4B), without change in fasting carbohydrate and fatty acid oxidation (Fig. 4E and F).

During the postprandial period, there was a significant decrease in energy expenditure (Fig. 4C) and in protein oxidation (Fig. 4D). Carbohydrate oxidation (Fig. 4E) was also reduced during the 1st h after meal intake, but the total fatty acid oxidation rate was not changed throughout the postprandial period (Fig. 4F). In contrast, a significant change occurred in the dynamics of DFA oxidation, with early postprandial increase and late postprandial reduction in the DFA oxidation rate after surgery (postprandial time × pre/postsurgery interaction, P < 0.0001) (Fig. 4G). There was no difference in postprandial lipid metabolism between the two types of bariatric operations (Supplementary Figs. 1 and 2). Fasting and postprandial β-hydroxybutyrate levels were elevated approximately twofold after bariatric surgery (Fig. 4H), suggesting an increase in hepatic fatty acid oxidation rate.

There was a significant improvement in hepatic Si within 12 days after surgery, as measured by fasting EGP in relation to insulin concentration or by the postprandial AUC of EGP (Table 1). This was associated with lower fasting glucose, insulin, and C-peptide levels after surgery. The percentage reduction in hepatic insulin resistance was strongly correlated with increased intra-abdominal fat storage (r = −0.79, P = 0.05). The percentage reduction in HOMA-insulin resistance was correlated with reduced DFA spillover (r = 0.76, P = 0.01). However, there was no improvement in whole-body Si calculated using the oral minimal model, likely reflecting the absence of improvement in peripheral Si. There were, however, two different patterns in glucose and insulin postprandial responses according to the type of surgery performed (Supplementary Fig. 3). The glucose incremental AUC increased nonsignificantly after SG, whereas it decreased in the three participants after BPD-DS (Table 1 and Supplementary Fig. 3). The rate of absorption of meal glucose was also faster after SG compared with before surgery, although total meal glucose absorption (AUC of meal Ra_glucose) did not change after SG. In contrast, the postprandial AUC of the meal Ra_glucose was consistently reduced after BPD-DS. The adipose tissue insulin resistance index (fasting Ra_NEFA × fasting insulin) was not significantly different after bariatric surgery (Table 1). Fasting leptin and total adiponectin were significantly decreased after bariatric surgery (Table 1).

We assessed the effects of the two types of bariatric operations on postprandial gut hormone concentrations (Fig. 5). After SG, there was an early postprandial increase in glucagon, GLP-1, GIP, and PYY levels versus before surgery. After BPD-DS, there was no difference in postprandial glucagon concentration. There was, however, a prolonged increase in postprandial GLP-1 and PYY responses but a blunted postprandial GIP response (Fig. 5).

We found that within 12 days after surgery, there was a significant decrease in cardiac DFA partitioning and in the EGP rate, along with a decrease in systemic DFA spillover and an increase in intra-abdominal adipose tissue DFA uptake. Enhanced uptake of DFA into adipose tissues and reduced systemic DFA spillover was also supported by a change in the pattern of the DFA oxidation rate, from a low early/high late to a high early/low late postprandial DFA oxidation rate. This happened in the absence of decreased gastrointestinal absorption of DFA, similar postprandial CM- and VLDL-TG levels, and in the absence of significant improvement in whole-body Si. These results point toward an early improvement in the storage capacity of DFA in intra-abdominal adipose tissues, with concomitant reduction of systemic DFA spillover. There was no evidence of decreased in DFA absorption and direct cardiac uptake of DFA from CM-TG, further supporting the reduction of adipose tissue DFA spillover as the main mechanism for reduced cardiac DFA partitioning with bariatric surgery in patients with T2D. However, it is possible that the increased postprandial endogenous Ra_NEFA after surgery may have competed with DFA for cardiac uptake as a possible explanation for the observed decrease in cardiac DFA uptake, without a net change in myocardial fatty acid utilization. The increase in β-hydroxybutyrate could also partly explain decreased cardiac DFA partitioning because uptake of ketone bodies by the myocardium is in proportion to their plasma concentration, competing with glucose and fatty acid oxidation (26,27).

The absence of evidence for intestinal fat malabsorption in patients with BPD-DS was surprising but was shown using two independent methods (oral [¹⁸F]-FTHA and oral [U¹³C]palmitate). The test meal that was served in the current study contained 50% carbohydrates, 30% fat, and 20% proteins but was hypocaloric, with only 9.4 g of fat and ∼358 kcal. Caloric restriction is common in the first weeks after bariatric surgery (3,6). Therefore, our experimental conditions reflect the usual postoperative meal intake in these patients. It is likely that intestinal fat malabsorption would be evident at higher dietary fat intake. Our results, however, do not suggest that impaired intestinal fat absorption is an important mechanism for the early metabolic improvements observed after bariatric surgery in patients with T2D. The same trends observed in postprandial DFA metabolism irrespective of the type of bariatric surgery (i.e., SG vs. BPD-DS) also support this interpretation. A standard liquid meal was used in the current study. More studies are needed to determine whether solid meals may change these results.

We confirmed the rapid improvement of hepatic Si and the absence of improvement of peripheral Si that were previously demonstrated in the first few days after bariatric surgery (5–7,9). We found an accelerated but quantitatively similar absorption of meal glucose after SG, as previously documented after significant weight loss after SG (28). However, we report for the first time a consistent decrease in meal glucose absorption after BPD-DS. Such lower meal glucose availability in circulation is associated with reduced postprandial insulin excursion and likely partly explains the superior short-term efficacy of BPD-DS to induce T2D remission compared with other bariatric surgical procedures (6). As previously described shortly after bariatric surgery (3,4), we found a major postprandial increase in GLP-1 and PYY and a decrease in circulating leptin levels with both SG and BPD-DS, which may contribute to improvement in adipose tissue fatty acid storage (29) and in postprandial TG and NEFA metabolism (30). Increased postprandial GIP after SG may also contribute to adipose tissue fatty acid uptake (31). Also, fasting acylated ghrelin was lower after bariatric surgery. This hormone was shown to stimulate muscle fatty acid oxidation, adipose tissue storage, and inhibition of lipolysis when acutely or chronically increased (32,33). Thus, reduced acylated ghrelin is not a likely explanation for the increase in adipose tissue DFA storage and reduced spillover we observed after bariatric surgery.

Fatty acids contribute for 50%–70% of ATP production in the healthy heart and to larger proportions in people with obesity or T2D (34). Although fatty acid availability is necessary for the maintenance of cardiac function, this greater reliance on fatty acids may also lead to a myocardial lipotoxicity that contributes to myocardial dysfunction over time (34–36). Fatty acids are supplied to the myocardium through the NEFA, CM-TG, and VLDL-TG pools (36). Approximately 2% of DFAs are used in the heart of healthy subjects, a proportion that is doubled in prediabetes (14). This increase in cardiac DFA partitioning is associated with increased myocardial oxidative metabolism, reduced left ventricular function, and reduced DFA uptake in abdominal subcutaneous and visceral adipose tissues in individuals with impaired glucose tolerance (15,21). We have shown that a 1-year lifestyle intervention leading to modest weight loss (−3.7 kg, 4% weight loss) in individuals with impaired glucose tolerance reduces cardiac DFA partitioning and myocardial oxidative metabolism and increases left ventricular function and abdominal adipose tissue DFA storage (19). However, a 7-day calorie- and saturated fat-restriction diet leading to only 1% weight loss actually increased cardiac DFA partitioning and reduced left ventricular function (20), suggesting that caloric restriction per se is not responsible for the observed reduction in cardiac DFA partitioning with lifestyle- or bariatric surgery-induced weight loss.

Cardiac fasting NEFA uptake was previously measured using intravenous [¹⁸F]-FTHA and PET imaging 6 months after bariatric surgery (Roux-in-Y gastric bypass or SG) with major weight loss (−23% total body weight) (37). The latter study showed reduced cardiac NEFA uptake after bariatric surgery, irrespective of the presence of T2D. This was associated with improved insulin sensitivity, increased insulin-mediated cardiac glucose uptake, and reduced left ventricular hypertrophy, output, and work load (37–39). After 6 weeks of a very low-calorie diet (550 kcal/day), a caloric intake similar to the average intake within the first 2 weeks after bariatric surgery (3), obese participants without T2D had decreased left ventricular mass, output, and work along with a decrease in myocardial NEFA uptake (40). From the latter studies, reduced cardiac hypertrophy and work and/or improved cardiac glucose metabolism are reasonable explanations for the reduction in cardiac NEFA uptake after bariatric surgery. We did not measure left ventricular function because of logistical constraints and, therefore, cannot rule out some cardiac remodelling to explain changes in cardiac DFA partitioning in our patients.

Our results are consistent with those of previous studies showing enhanced adipose tissue DFA storage after weight loss (19,41). In contrast, some previous studies did not observe changes in subcutaneous or visceral adipose tissue NEFA uptake per volume of tissue after surgery (37,39) or after diet-induced weight loss (42). This is consistent with different regulation of adipose tissue fatty acid storage from the NEFA and the CM-TG pools. Our results are also consistent with a predominant role of intra-abdominal adipose tissues as a source for systemic DFA spillover (43,44). We previously showed reduced NEFA spillover from intravenous lipid emulsion during euglycemic-hyperinsulinemic clamp conditions in subjects with T2D 3 days, but not 3 or 12 months after BPD-DS (5). NEFA spillover from intravenous administration of TGs during continuous oral feeding did not change in obese subjects with T2D after a 14% weight loss induced by lifestyle changes (45). The role of enhanced adipose tissue DFA spillover in obesity and T2D is very controversial (46). Nevertheless, our results suggest rapid and dynamic adaptations of adipose tissues DFA storage after bariatric surgery, at least leading to transient reduction in systemic DFA availability to other organs. We found a strong association of the reduction in hepatic insulin resistance with increased intra-abdominal fat storage and with reduced DFA spillover, supporting a close link between intra-abdominal adipose tissue DFA spillover and hepatic insulin resistance in T2D.

We found no difference in hepatic DFA uptake or partitioning. Hepatic DFA uptake and partitioning results from the integration of DFA uptake from CM remnants and from recirculation of DFA in the NEFA pool minus secretion of DFA in VLDL-TG (18). We, however, observed an increase in plasma β-hydroxybutyrate levels, suggesting increased hepatic fatty acid oxidation. Of note, any increase in hepatic DFA oxidation would not be expected to change liver DFA uptake from [¹⁸F]-FTHA PET imaging because this tracer is trapped into mitochondria and esterified into lipid pathways once taken up by tissues (47). Hepatic NEFA uptake per volume of tissue was previously found to be unchanged in obese subjects without or with T2D 6 months after bariatric surgery (38). We observed no change in postprandial VLDL concentrations and no change in [U-¹³C]palmitate from the liquid meal into VLDL-TG, arguing also against a major change in hepatic DFA recirculation into VLDL during the first 6-h postprandial period within 12 days after bariatric surgery. To our knowledge, no study has measured the VLDL-TG secretion rate this early after bariatric surgery. However, the VLDL-apoB100 secretion rate was decreased 6 months after SG along with an increase in the VLDL-apoB100 catabolic rate during continuous oral feeding in obese individuals without T2D (48).

The 8- to 12-day period was the earliest possible time to perform our postsurgical measurements to limit the confounding effects of weight loss. However, the participants already displayed a 5% weight loss that likely explained at least part of the improvement we observed in cardiac and intra-abdominal DFA partitioning (19). The number of subjects included is also limited and include two different bariatric surgical procedures because of the logistical and cost challenges of performing such complex studies in this population early after surgery. Although there were clear differences in postprandial glucose and hormonal excursion between SG and BPD-DS, the same trends were observed in DFA metabolism, endogenous glucose appearance, Si, whole-body substrate oxidation, and weight loss in all of the subjects from the two groups. Another limitation of the current study is the absence of quantification of gastric emptying rate.

We conclude that bariatric surgery rapidly corrects increased cardiac DFA partitioning and reduced intra-abdominal adipose tissue DFA uptake in T2D, with a significant reduction in systemic DFA spillover and hepatic insulin resistance. More studies are needed to determine the impact of these metabolic changes on cardiac structure and function and to determine whether these changes are sustained over time.

Acknowledgments. The authors acknowledge the contribution of surgeons, nurses, and the medical team of the bariatric surgery program at IUCPQ. The coinvestigators and collaborators of the REMISSION study are (in alphabetical order): C. Bégin, L. Biertho., M. Bouvier, S. Biron, P. Cani, A.C. Carpentier, A. Dagher, F. Dubé, A. Fergusson, S. Fulton, F.S. Hould, F. Julien, T. Kieffer, B. Laferrère, A. Lafortune, S. Lebel, O. Lescelleur, E. Levy, A. Marette, S. Marceau, F. Picard, P. Poirier, D. Richard, J. Schertzer, A. Tchernof, and M.C. Vohl. The authors would like to thank Lucie Bouffard, Mélanie Fortin, Maude Gérard, Diane Lessard, Caroll-Lynn Thibodeau (from the Division of Endocrinology, Department of Medicine, Centre de recherche du CHU de Sherbrooke, Université de Sherbrooke, Sherbrooke, Québec, Canada), and Éric Lavallée (from the Department of Nuclear Medicine and Radiobiology, Centre de recherche du CHU de Sherbrooke, Université de Sherbrooke, Québec, Canada) for their technical support.

Funding and Duality of Interest. This work was supported by Canadian Institutes of Health Research (CIHR) operating grants 341582 and 299962, CIHR Team (REMISSION) grant on bariatric care TB2-138776, a Diabetes Canada operating grant 3-14-4507-AC, and an investigator-initiated study grant from Johnson & Johnson Medical Companies (grant ETH-14-610). A.-M.C. was supported by Diabetes Canada and Fonds de Recherche du Québec - Santé postdoctoral fellowships. L.B. and A.T. are codirectors of the Research Chair in Bariatric and Metabolic Surgery at Laval University. A.C.C. holds the Canada Research Chair in Molecular Imaging of Diabetes. L.B. and A.T. also receive research funding from Medtronic for studies unrelated to the present manuscript. No other potential conflicts of interest relevant to this article were reported.

Funding sources for the trial had no role in the design, conduct, or management of the study; in collection, analysis, or interpretation of data; or in the preparation of the present manuscript and decision to publish.

Author Contributions. A.-M.C., C.N., D.P.B., F.F., M.N., M.P., and S.P. performed experiments. A.-M.C., C.N., F.F., and E.E.T. analyzed data. A.-M.C., S.C.C., B.G., E.E.T., S.L., L.B., A.T., and A.C.C. interpreted results of experiments. A.-M.C. and A.C.C. drafted the manuscript. A.C.C. contributed to the conception and design of research. All authors edited and revised the manuscript and approved the final version of the manuscript. A.C.C. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in abstract form at the 78th Scientific Sessions of the American Diabetes Association, Orlando, FL, 22–26 June 2018.