Article Text

Effect of weight-maintaining ketogenic diet on glycemic control and insulin sensitivity in obese T2D subjects
  1. Aurora Merovci1,
  2. Brittany Finley1,
  3. Andrea Hansis-Diarte1,
  4. Sivaram Neppala1,
  5. Muhammad A Abdul-Ghani2,
  6. Eugenio Cersosimo3,
  7. Curtis Triplitt1,
  8. Ralph A DeFronzo1
  1. 1Department of Medicine, Diabetes Division, University of Texas Health Science Center, San Antonio, Texas, USA
  2. 2Diabetes Division, University of Texas Health Science Center, San Antonio, Texas, USA
  3. 3Medicine, Texas Diabetes Institute, San Antonio, Texas, USA
  1. Correspondence to Professor Ralph A DeFronzo; defronzo{at}uthscsa.edu

Abstract

Introduction Low carbohydrate ketogenic diets have received renewed interest for the treatment of obesity and type 2 diabetes. These diets promote weight loss, improve glycemic control, and reduce insulin resistance. However, whether the improvements in glycemic control and insulin sensitivity are secondary to the weight loss or result from a direct effect of hyperketonemia is controversial.

Research design and methods 29 overweight obese subjects were randomized to one of three dietary interventions for 10 days: (1) Weight-maintaining standard diet; (2) Weight-maintaining ketogenic diet; (3) Weight-maintaining ketogenic diet plus supplementation with the ketone ester of beta-hydroxybutyrate (β-OH-B), 8 g every 8 hours. At baseline, all subjects had oral glucose tolerance test, 2-step euglycemic insulin clamp (20 mU/m2.min and 60 mU/m2.min) with titrated glucose and indirect calorimetry.

Results Body weight, fat content, and per cent body fat (DEXA) remained constant over the 10-day dietary intervention period in all three groups. Plasma β-OH-B concentration increased twofold, while carbohydrate oxidation decreased, and lipid oxidation increased demonstrating the expected shifts in substrate metabolism with institution of the ketogenic diet. Glucose tolerance either decreased slightly or remained unchanged in the two ketogenic diet groups. Whole body (muscle), liver, and adipose tissue sensitivity to insulin remained unchanged in all 3 groups, as did the plasma lipid profile and blood pressure.

Conclusion In the absence of weight loss, a low carbohydrate ketogenic diet has no beneficial effect on glucose tolerance, insulin sensitivity, or other metabolic parameters.

  • Ketones
  • Diabetes Mellitus, Type 2
  • Diet
  • Obesity

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

WHAT IS ALREADY KNOWN ON THIS TOPIC

  • Ketogenic diets promote weight loss, improve glycemic control, and reduce insulin resistance.

WHAT THIS STUDY ADDS

  • This study demonstrates that, in the absence of weight loss, the ketogenic diet does not improve glycemic control, muscle, or liver insulin sensitivity.

  • However, it does lead to a significant increase in insulin secretion.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

  • The study offers a novel perspective on the metabolic effects of short-term ketogenic diet intervention, particularly regarding weight loss.

  • It further supports the fact that during ketosis, weight loss is a crucial factor in improving glycemic control.

  • However, it also notes that improvements in insulin secretory capacity can occur even without weight loss.

Introduction

Type 2 diabetes (T2D) and obesity have become global healthcare problems. Approximately 422 million people have T2D worldwide,1 and more than 37 million Americans have diabetes.2 The diabetes epidemic is being driven by the obesity epidemic.3 In USA, 42% of the adult population is obese (body mass index (BMI) >30 kg/m2) and 30% is overweight.4 Obesity is an insulin-resistant state,5–7 and in patients with T2D obesity exacerbates the underlying insulin resistance (IR) and promotes progressive β cell failure.8 9 Although the newer antihyperglycemic agents have potent appetite-suppressive effect and weight loss promoting effect,10 11 the mainstay of therapy remains lifestyle intervention.12 A number of diets (low carbohydrate ketogenic diet, Mediterranean diet, high fiber diet, DASH) have been employed to promote weight loss with variable success.12 Most recently, renewed interest in a hypocaloric, low carbohydrate ketogenic diet has emerged as an effective strategy to promote weight loss in obese individuals and to improve glycemic control in patients with T2D.13–23 The beneficial effects of a ketogenic diet on glycemic control, insulin sensitivity and β cell function in obese patients with T2D variably has been attributed to weight loss or to hyperketonemia per se.24–26 To examine this controversy, in the present study we determined the effect of a ketogenic diet on glucose metabolism in people with T2D while maintaining a constant body weight. Our results demonstrate, for the first time, that a low carbohydrate ketogenic diet with or without keto-ester supplementation in obese T2D individuals does not improve glycemic control, insulin sensitivity, or β cell function in the absence of weight loss.

Research design and methods

Subjects

Twenty-nine overweight/obese (BMI=27.5–42 kg.m²) subjects with T2D, ranging in age from 18 years to 70 years and HbA1c from 7.0% to 10.5%, participated in three study protocols: (1) Standard, weight-maintaining diet;2 (2) Low carbohydrate, ketogenic diet;3 (3) Low carbohydrate, ketogenic diet with keto ester supplementation. Other than diabetes, subjects were in good general health based on medical history, physical exam, routine chemistry and hematology tests, urinalysis, and EKG. Subjects with diabetes were drug-naïve or were treated with metformin, sulfonylurea, or dipeptidyl peptidase-4 inhibitor. Subjects taking insulin, pioglitazone, SGLT2i, or GLP-1 receptor agonist were excluded. Subjects taking any medication known to affect glucose metabolism (other than antidiabetic agents) were excluded. Body weight was stable (±4 pounds) over the preceding 3 months. Subjects who participated in an excessively heavy exercise program were excluded.

Following completion of baseline studies, subjects were randomized into three groups: (1) Standard, weight-maintaining diet containing 25%–35% protein, 45%–55% carbohydrate, 20%–30% fat; (2) Weight-maintaining, ketogenic diet containing 15%–25% protein, 5%–10% carbohydrate, 70%–80% fat; (3) Weight-maintaining, diet containing 15%–25% protein, 5%–10% carbohydrate, 70%–80% fat supplemented with the ketone ester of beta-hydroxybutyrate (β-OH-B) (HVMN, San Francisco, California, USA), 8 grams every 8 hours, to further increase the plasma ketone concentration.27 Antidiabetic medications were not changed during the study in any participant. The duration of diet intervention was 10 days which, based on the study by Stubbs et al,28 should be sufficient to observe the effect of hyperketonemia on glucose/lipid metabolism and energy expenditure.

Study procedures: interventions

Diet

The diets consisted of a 10-day rotating menu using NutriAdmin for meal planning and Nutrition Maker for analysis of recipes. Food was prepared in the metabolic kitchen at the Texas Diabetes Institute by a certified dietician. Subjects reported to the diet kitchen at 08:00 hours Monday through Saturday where they ate their breakfast and picked up food for their lunch and dinner. On Saturday, the participants were given food for their Sunday meals also. Subjects maintained a daily dietary log in which they recorded all food consumed. On each visit the subjects were weighed, the dietary log was reviewed, and the caloric content of the diet was adjusted to maintain the body weight constant within 2 pounds. Blood pressure was measured in the reclining position on every study visit. Blood samples were obtained every day to measure the plasma concentrations of β-OH-B, glucose, insulin, C peptide, and glucagon. During the 10-day dietary treatment period, participants were instructed to consume only the diet prepared in the metabolic kitchen, with the last meal finished no later than 22:00 hours. Subjects were requested to drink only water outside this time window.

Ketone ester of β-OH-B supplement

In addition to a keto diet, participants in group 3 were supplemented with 8 g ketone ester of β-OH-B (HVMN, San Francisco, California, USA) every 8 hours in the form of a drink and were requested to finish the ester intake no later than 22:00 hours following their meal. An 8-hour interval between beta hydroxybutyrate (BOHB) ingestion was chosen to enhance compliance with the treatment regimen. The total calorie load of ketone ester per day was 122 calories. The low carbohydrate ketogenic diet was well tolerated by all subjects.

Indirect calorimetry

During the screening visit, whole body substrate oxidation rates and energy expenditure were measured with indirect calorimetry29 at 08:00 hours after a 10-hour overnight fast, with a ventilated hood system for 60 min. For each subject, his/her individual weight maintenance energy requirement was estimated as 1.5 times the resting energy expenditure, obtained by indirect calorimetry. Repeat indirect calorimetry test was performed on day 8 of treatment in the fasting state before the oral glucose tolerance test (OGTT). Indirect calorimetry also was performed during each insulin clamp study (see below).

Oral glucose tolerance test

All subjects received 75 g OGTT at 08:00 hours following a 10-hour overnight fast to document glucose tolerance before diet intervention. OGTT was repeated on day 8 after initiation of the diet intervention. Blood samples were drawn at −30 min,–15 min and 0 min and every 15 min thereafter for 2 hours for determination of plasma glucose, insulin, C peptide, glucagon, and β-OH-B concentrations. Baseline OGTT and screening visit were conducted approximately 2–3 weeks before starting diet treatment, while the baseline two-step euglycemic clamp was performed approximately 3 days after the OGTT.

DEXA scan

A DEXA scan was performed during the screening visit and on day 8 of the diet intervention to determine total body fat content, lean mass, and percentage body fat.

Fibroscan

Fibroscan of the liver was performed during the screening visit in the fasting state and on day 8 after the start of dietary treatment to measure liver fat and fibrosis.30

Two-step euglycemic insulin clamp

A two-step euglycemic insulin clamp was performed twice, at baseline and on day 10 of the diet treatment. At baseline all participants underwent a two-step euglycemic insulin clamp (20 mU/m2⋅ min and 60 mU/m2⋅ min) to raise plasma insulin concentration by ∼30–40 μU/mL and ∼70–80 μU/mL, respectively;31 each step lasted 90 min. The rate endogenous (primarily reflects hepatic glucose production (HGP)) and total glucose disposal (TGD) or glucose disposal rate were measured during the 80–90 min time period of the first insulin clamp step and during the 170–180 min time period of the second insulin clamp step during which the plasma glucose and insulin concentration were at a steady state. At 06:30 hours, after 10 hours overnight fast, participants reported to the Texas Diabetes Institute CRC. A catheter was placed into an antecubital vein for infusion of all test substances, and a second catheter was inserted retrogradely into a vein on the dorsum of the hand, which was inserted into a heated box (60°C) to obtain arterialized blood. A 3-3H-glucose infusion (prime=40 µCi × (fasting plasma glucose (FPG))/100)); continuous infusion=0.40 µCi/min) was initiated 180 min prior to the start of insulin and continued throughout the 180 min clamp procedure to measure total body (primarily reflects muscle), hepatic, and adipose tissue insulin sensitivities.32 Indirect calorimetry was performed at baseline (−45 min to 0 min) before start of insulin infusion and from 60–90 min and 150–180 min after the start of insulin. Plasma samples for determination of titrated glucose specific activity were obtained at −180 min, –30 min, −20 min, –10 min, −5 min, and 0 min and every 10–15 min during the two-step insulin clamp. After the start of insulin, the plasma glucose concentration was allowed to decrease to 100 mg/dL at which time a variable infusion of 20% dextrose was initiated to maintain plasma glucose level at ~100 mg/dL.

Analytical techniques

Plasma glucose was measured by the glucose oxidase reaction (Glucose Oxidase Analyzer; Analox, Instruments, Stourbridge, UK). Plasma insulin concentration was measured by the IRMA method, DIA source immunoassays S.A., Nivelles, Belgium). Plasma C peptide concentration was measured by the IRMA method (MP Biomedicals, Irvine, California, USA). Plasma β-OH-B concentration was measured by the colorimetric method (Millipore Sigma, Burlington, Massachusetts, USA). Plasma 3-3H-glucose radioactivity was measured in Somogyi precipitates.

Calculations and statistical analysis

The incremental area under the plasma glucose, insulin, C peptide and B-HO-B concentration curves was calculated with the trapezoid rule. Insulin secretion was calculated as the ratio between the incremental area under the plasma C peptide concentration to the incremental area under the plasma glucose concentration during the OGTT (ΔC-Pep0–120/ΔG0–120). The β cell function was measured as the insulin secretion/IR or Disposition Index, where IR was the inverse of TGD measured with the two-step euglycemic insulin clamp (ΔC-Pep0–120/ΔG0–120 ÷ IR). Homeostasis model assessment (HOMA)-IR during days 3, 5, 8 was calculated as: (plasma insulin × plasma glucose)/405 with glucose in mg/dL and insulin in mU/L. Insulin Sensitivity Index (M/I) comparison in both euglycemic steps was done with paired t-test. Insulin clearance rate was calculated during the first insulin clamp step as follows: 20 mU/m2⋅min/(SSPI80–90−FPI) and for the second step 60 mU/m2⋅ min/(SSPI170–180−FPI).

Under steady-state postabsorptive conditions, the basal rate of endogenous glucose appearance (Ra) equaled the 3-3H-glucose infusion rate divided by steady-state plasma tritiated glucose specific activity. During the insulin clamp, non-steady conditions for 3-3H-glucose specific activity occur and the rate of glucose appearance (Ra) was calculated using the Steele equation. The rate of residual EGP during the insulin clamp was calculated by subtracting the exogenous glucose infusion rate from the tracer-derived Ra. The insulin-stimulated rate of TGD was calculated by adding the rate of residual EGP to the exogenous glucose infusion rate.

Data are presented as mean±SEM. Changes from baseline (before vs after treatment) in TGD, EGP, ΔC-Pep0–120/ΔG0–120, and ΔC-Pep0–120/ΔG0–120 divided by IR within each group were compared with paired t-test. The repeated measures analysis of variance (ANOVA) was used to test the main effects and interactions of measurements across groups and ANOVA with Bonferroni post hoc correction was used to analyze the differences between the groups before and after treatment, using SPSS statistics V.29.0 (IBM). Statistical significance was considered at α<0.05.

Results

Clinical, laboratory, and anthropometric characteristics

In all the body weight of 29 subjects changed by less than 2 pounds over the 10-day dietary intervention period and end-of-study weight was not significantly different from the baseline value (table 1; figure 1A). Fat mass, per cent body fat, and lean mass did not change significantly in any of the three groups (table 2). Liver fat (CAP) and liver fibrosis score (kPa), determined by Fibroscan, did not change with any of the three dietary interventions (table 2).

Figure 1

Effect of standard diet (n=8), keto diet (n=10), and keto diet plus keto esters (n=11) on body weight (A), fasting plasma beta-hydroxybutyrate concentration (C). * p<0.01–0.001 versus baseline.

Table 1

Clinical, laboratory, and anthropometric characteristics

Table 2

Effect of dietary intervention on fat mass, per cent body fat, and lean mass, liver steatosis (CAP) and fibrosis (kPa)

In both the keto diet and keto diet plus keto ester groups the fasting plasma β-OH-B concentration rose significantly by day 3 and remained elevated throughout the 10-day treatment period (figure 1B). The diet-induced increase in plasma ketone concentration was ~twofold above the baseline (figure 1B), from 0.22 mM (baseline) to 0.44 mM (mean of days 3–10 postdiet) (p<0.001) and was similar in the keto diet and keto diet plus keto ester groups. We attribute the similar rise in plasma B-OH-B in the two keto diet groups to the rapid metabolism of ketone bodies.27 28 33 34 To confirm this, we measured the plasma B-OH-B concentration every 15 min following ingestion of 8 grams of the ketone ester of β-OH-B in three obese patients with T2D (online supplemental figure 4). The fasting B-OH-B concentration (0.23 mM) rose to a peak of 0.57 mM at 45 min and returned to the baseline value after 120 min. The incremental area under the fasting plasma β-OH-B concentration over the 10-day treatment period was similar between the two ketogenic diet groups: keto diet group and keto diet plus ester group. However, there was a significant difference between keto diet and keto diet plus keto esters versus standard diet group (ANOVA, p=0.016).

Supplemental material

Supplemental material

Ten days of standard diet treatment (group 1) had no effect on fasting respiratory quotient (RQ), carbohydrate oxidation, or fat oxidation (table 3; figure 2). In both the keto diet and keto diet plus keto ester groups the RQ and carbohydrate oxidation decreased significantly, while fat oxidation increased significantly (table 3; figure 2). These changes in carbohydrate and fat oxidation and RQ, coupled with increase in plasma β-OH-B concentration, demonstrate that the ketogenic diet was effective and produced a significant switch in substrate metabolism.

Figure 2

Glucose oxidation (GOX) and fat oxidation (FOX) prediet and postdiet in the standard diet (n=8), keto diet (n=10), and keto diet plus ester (n=11) groups. *p<0.01, **p<0.001.

Table 3

Oral glucose tolerance test and indirect calorimetry

Fasting glucose, HbA1c, fructosamine

FPG concentration declined modestly by ~15–20 mg/dL in all three groups (p=NS between groups) (figure 1C). There was no significant change both between and within the groups, which suggests that diet intervention did not have a measurable effect on FPG levels. Since the initial OGTT visit and insulin clamp test were performed on different days at a 10-day interval, it is not surprising that there was some small non-significant difference in the FPG concentration measured on the 2 days.

Neither the plasma fructosamine concentration nor the HbA1c changed significantly from baseline to day 10 in any group (table 1).

Oral glucose tolerance test

On day 8 of the dietary intervention the plasma glucose concentration during the OGTT was similar to that at baseline in the standard diet group and in the keto diet plus keto ester groups (figure 3A). In the keto diet group there was a small but significant deterioration in glucose tolerance on day 8 versus baseline (figure 3A).

Figure 3

Plasma glucose (A), insulin (B), and C peptide (C) concentrations during OGTT prediet treatment and postdiet treatment in the standard diet (n=8), keto diet (n=10), and keto diet plus keto ester (n=11) groups. *p<0.05–0.01. OGTT, oral glucose tolerance test.

During the OGTT, plasma insulin (figure 3B) and C peptide (figure 3C) concentrations were significantly higher in the keto diet and keto diet plus keto ester groups on day 8 versus baseline (p=NS vs standard diet) (figure 3B,C). In the standard diet group, the plasma insulin and C peptide concentrations during the OGTT at day 8 were slightly but not significantly higher than baseline. Insulin secretion, measured as the plasma C peptide AUC (0–120 min), increased significantly in the keto diet group and remained unchanged in keto diet plus keto esters group and in the standard diet group (table 4). The disposition index, measured as C peptide AUC (0–120)/glucose AUC (0–120) divided by IR (where IR=TGD during insulin clamp) remained unchanged in all groups (table 4).

Table 4

Effect of standard diet, keto diet, and keto diet plus keto esters on parameters of insulin secretion

Fasting plasma free fatty acid (FFA) concentration was similar in all three groups at baseline and did not change with any of the three dietary interventions (online supplemental figure 1). Suppression of plasma FFA during the OGTT and during both steps of the insulin clamp was similar postdiet versus prediet in all three groups (online supplemental figure 1). HOMA-IR tended to decrease in the keto diet group but was unchanged in the keto diet plus keto esters group (table 4). The repeated measures ANOVA was performed to identify the interaction effect of HOMA-IR (days 3, 5, and 8) by group. There was no interaction for changes in HOMA-IR when comparing the diet groups (F2 26 = 0.69, p=0.933). The HOMA-IR did not change significantly before versus after 10 days of diet treatment in three diet treatment groups, and there were no significant differences observed between the groups (table 4). The repeated measures ANOVA was performed to identify the interaction effect of Matsuda Index before versus after diet treatment by group. There was no interaction for changes in Matsuda Index when comparing the diet groups (F2 26 = 1.073, p=0.309). The Matsuda Index did not change significantly before versus after 10 days of diet treatment in three diet treatment groups, and there were no significant differences observed between the groups (table 4).

Insulin clamp: hepatic and muscle insulin sensitivity

The steady state plasma insulin concentration during step 1 of the insulin clamp was ~35–38 uU/mL and during step 2 was ~70–80 uU/mL, similar in all three groups at baseline and after 10 days of dietary treatment (online supplemental figure 2). The steady state plasma glucose concentration during the insulin clamp was achieved at time 70 min in the standard diet group and at 115 min in the keto diet and keto diet plus keto esters diet groups and remained constant thereafter for the duration of the insulin clamp (table 5 and online supplemental figure 3). Baseline HGP was similar in all three groups prior to dietary treatment and was unaltered after 10 days of keto diet, keto diet plus keto esters, and standard diet therapy (figure 4A). During the first insulin clamp step, total body glucose disposal increased slightly (p=NS) and similarly in all three groups in the prediet and postdiet studies. During the second insulin clamp, total body glucose uptake increased significantly (p<0.001) and similarly in all three groups. During the repeat insulin clamp on day 10, insulin-stimulated glucose uptake during both insulin clamp steps was similar to the baseline study prior to dietary intervention in all three groups (figure 4B).

Figure 4

Hepatic glucose production (A) and total body (muscle) glucose disposal (B) at baseline and during step 1 and step 2 of the insulin clamp in the standard diet (n=8), keto diet (n=10), and keto diet plus keto esters (n=11) groups. No significant differences were observed between prediet and postdiet for any parameter. p<0.05–0.01.

Table 5

Steady-state PGC and PIC during a two-step euglycemic clamp before and after treatment

M/I remained unchanged after treatment in the three diet treatment groups. There was no change in the insulin clearance rate in either the ketogenic diet or ketogenic diet plus keto esters supplement groups during both steps of euglycemic insulin clamp (table 5).

Lipid profile, blood pressure

Systolic blood pressure decreased modestly during the standard diet treatment and did not change in the keto diet and keto diet plus esters groups (table 1). No significant changes in diastolic blood pressure, heart rate, total cholesterol, LDL cholesterol, HDL cholesterol, or triglycerides were observed in any of the three dietary intervention groups (table 1).

Calorie intake during 10 days of diet treatment

The mean daily caloric intake required to maintain a stable body weight was 2549±133, 2756±167, and 3165±133 in the standard diet, keto diet and keto diet plus keto ester groups. The estimated energy expenditure at baseline was higher in the keto diet plus keto ester group compared with the keto diet group which was higher than that in the standard diet group.

Adverse events

One subject reported mild nausea when ingesting the keto ester supplement but consumed all three daily supplements and completed the entire study. Otherwise, there were no adverse events.

Discussion

Multiple studies13–19 and meta-analyses/reviews20–23 have demonstrated that a low carbohydrate, ketogenic diet can enhance glycemic control in obese individuals with T2D. However, there remains controversy regarding whether the favorable effects of the ketogenic diet on glycemia are secondary to the weight loss or result from a specific effect of the hyperketonemia on glucose metabolism. To address this question, we examined in obese patients with T2D the effect of a ketogenic diet alone or in combination with a keto ester supplement on glucose homeostasis and insulin sensitivity while maintaining a constant body weight.

Our findings, for the first time, demonstrate that a 10-day low carbohydrate ketogenic diet, without concomitant weight loss, has no effect on glycemic control or on muscle, hepatic or adipose tissue sensitivity to insulin. Additionally, we observed no effect of ketogenic diet on the plasma lipid profile and blood pressure in the absence of weight loss. To ensure compliance with the dietary regimen, participants reported every morning to the metabolic kitchen at the Texas Diabetes Institute where they ate their breakfast and picked up their individually prepared meals (lunch and dinner) for the day. Daily review of each subject’s dietary log indicated adherence with the dietary regimen. Furthermore, compliance was validated through indirect calorimetry, demonstrating a decrease in RQ, increase in lipid oxidation, and decrease in carbohydrate oxidation.

Previous studies that examined the effect of a ketogenic diet on glycemic control and insulin sensitivity in obese patients with T2D did not examine the postprandial glucose response and used HOMA-IR to provide an index of insulin sensitivity.33 34 HOMA-IR primarily reflects hepatic IR and is a static measure.35 Moreover, in all of these studies the institution of the ketogenic diet was associated with significant weight loss,33 34 and weight loss inherently improves glycemic control and IR.6 8 Our study used the gold standard euglycemic insulin clamp technique combined with tritiated glucose and measurement of plasma FFA concentration to assess tissue-specific insulin sensitivity (muscle, liver, adipose tissue) while maintaining constant body weight after initiation of the ketogenic diet. The results clearly demonstrate that, in the absence of weight loss, the ketogenic diet does not improve glycemic control or insulin sensitivity. Although there was a trend for HOMA-IR and the Matsuda Index of insulin sensitivity to worsen in the keto diet group, these changes did not reach statistical significance. Further, in the keto diet plus keto esters group, there was no change in HOMA-IR or the Matsuda Index. Most importantly, using the gold standard euglycemic insulin clamp with tritiated glucose there was absolutely no change in either hepatic or muscle sensitivity, which remained unchanged in all three groups. While gluconeogenesis was not directly measured, it also is unlikely to have changed, given that HGP remained unchanged.

Insulin-stimulated muscle glucose uptake (second insulin clamp step) and insulin-mediated suppression of HGP and plasma FFA concentration remained unchanged after 10 days of the standard diet and ketogenic diet, with or without keto ester supplementation.

A novel discovery in our study is the consistent increase in postglucose plasma insulin response during the OGTT following institution of the ketogenic diet alone or in combination with keto ester supplements. This increase in insulin secretion was paralleled by a similar rise in plasma C peptide concentration. In the standard diet group, neither the plasma insulin nor C peptide responses changed significantly and the fasting plasma ketone concentration remained unchanged throughout the dietary intervention period.

A compensatory increase in insulin secretion in response to IR is well established.9 36 However, no change in muscle, hepatic, or adipose tissue insulin sensitivity was observed during the insulin clamp in either of the ketogenic diet groups. Although the increase in plasma insulin and C peptide concentrations during the OGTT in the keto diet group could be explained, in part, by the increase in the plasma glucose concentration, hyperglycemia cannot explain the increase in plasma insulin and C peptide concentrations in the keto diet plus keto esters group. More likely, we believe that the stimulation of insulin secretion is the direct result of the elevated ketone levels on the β cell. Ketones are converted to AcCoA which are transported into and metabolized in mitochondria to generate ATP which is essential for the maintenance of normal β cell function.37 38 The effect of ketones on glucose-stimulated insulin secretion (GSIS) is controversial with some studies showing a stimulatory effect,37 39 while other studies have shown an inhibitory effect.40 These conflicting results may be explained by differences in the duration of exposure: acute ketone exposure producing an inhibitory effect on GSIS,40 while chronic exposure causing an enhancing effect on GSIS.37 39 Whatever the mechanism, our results demonstrate that both the keto diet and keto diet plus keto esters dietary intervention for 8–10 days augments GSIS.

While our study provides valuable insights about the effect of a ketogenic diet, it has some limitations. First, the duration of the ketogenic diet intervention (10 days) is relatively short. However, because ketones suppress the appetite,28 41 sustained long-term maintenance of body weight is difficult. Second, the sample size in each group is relatively small. However, measurement of insulin sensitivity with the insulin clamp is highly reproducible when the same subject is studied on repeat occasions.31 Since there was no increase in insulin sensitivity in muscle, liver, or adipose tissue in any of the 21 subjects treated with the keto diet or keto diet plus keto esters, we believe that increasing the sample size would be unlikely to have any effect on the results. It is possible that a longer period of carbohydrate restriction might yield more pronounced changes in plasma ketone levels and improvements in glycemic control and insulin sensitivity.

Conducting such a study would require participants to stay on a metabolic ward to closely monitor food intake to ensure constant body weight. However, recruitment would become a problem since participants may find it challenging to take time off from work for such an extended in-hospital stay. Moreover, the cost of the study would increase markedly due to the need for prolonged hospitalization and participant monitoring. Another limitation of the study is that the caloric intake was not adjusted on day 8 of the diet treatment to accommodate the intake of 75 grams of carbohydrates from the OGTT. However, the plasma ketone concentration remained unchanged and elevated on study days 9 and 10. Therefore, we do not believe that the OGTT affected the overall results of the study.

In summary, our study demonstrates for the first time that a 10-day ketogenic diet, with or without exogenous ketone supplementation, while maintaining constant body weight does not improve glucose tolerance, insulin sensitivity, plasma lipid profile, or blood pressure in obese patients with T2D. However, it does reveal a significant increase in pancreatic insulin secretion induced by dietary elevation in plasma ketone concentration.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.

Ethics statements

Patient consent for publication

Ethics approval

This study involves human participants and was approved by the UT Health San Antonio Institutional Review Board (IRB# HSC2000230H). Participants gave informed consent to participate in the study before taking part.

Acknowledgments

The authors thank Lorrie Albarado for her expert assistance in preparing the manuscript for publication. The authors also thank their CRC nurses (Anita Rodriguez BSN, James King, RN) for assistance in performing the insulin clamp and OGTT studies.

References

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Footnotes

  • Contributors The study was conceived by AM and RAD. AM performed the insulin clamp, OGTT, and follow-up studies with assistance from Anita Rodriguez BSN, Brittany Finley MS, RDN, Deena Murphey. AM wrote the first draft of the manuscript that was subsequently revised by RAD and then reviewed by all contributing authors. AM is the guarantor of this work and, as such, has 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.

  • Funding Funding was provided by 80 20 Foundation grant.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.