Exogenous ketosis in patients with type 2 diabetes: Safety, tolerability and effect on glycaemic control

Abstract Introduction Ketogenic diets have shown to improve glycaemic control in patients with type 2 diabetes. This study investigated the safety, tolerability, and effects on glycaemic control in patients with type 2 diabetes of an exogenous ketone monoester (KE) capable of inducing fasting‐like elevations in serum β‐hydroxybutyrate (βHB) without the need for caloric or carbohydrate restriction. Methods Twenty one participants (14 men and 7 women, aged 45 ± 11 years) with insulin‐independent type 2 diabetes, and unchanged hypoglycaemic medication for the previous 6 months, were recruited for this non‐randomised interventional study. Participants wore intermittent scanning glucose monitors (IS‐GM) for a total of 6 weeks and were given 25 ml of KE 3 times daily for 4 weeks. Serum electrolytes, acid‐base status, and βHB concentrations were measured weekly and cardiovascular risk markers were measured before and after the intervention. The primary endpoints were safety and tolerability, with the secondary endpoint being glycaemic control. Results The 21 participants consumed a total of 1,588 drinks (39.7 litres) of KE over the course of the intervention. Adverse reactions were mild and infrequent, including mild nausea, headache, and gastric discomfort following fewer than 0.5% of the drinks. Serum electrolyte concentrations, acid‐base status, and renal function remained normal throughout the study. Compared to baseline, exogenous ketosis induced a significant decrease in all glycaemic control markers, including fructosamine (335 ± 60 μmol/L to 290 ± 49 μmol/L, p < .01), HbA1c (61 ± 10 mmol/mol to 55 ± 9 mmol/mol [7.7 ± 0.9% to 7.2 ± 0.9%], p < .01), mean daily glucose (7.8 ± 1.4 mM to 7.4 ± 1.3 mM [140 ± 23 mg/dl to 133 ± 25 mg/dl], p < .01) and time in range (67 ± 11% to 69 ± 10%, p < .01). Conclusions Constant ketone monoester consumption over 1 month was safe, well tolerated, and improved glycaemic control in patients with type 2 diabetes.


| INTRODUC TI ON
Exogenous ketosis has therapeutic potential across a diverse range of chronic metabolic diseases, including epilepsy, neurodegenerative diseases, heart failure and diabetes. 1 Very low carbohydrate or ketogenic diets can reverse type 2 diabetes, 2 and exogenous ketosis without carbohydrate restriction lowers blood glucose. As examples, d-β-hydroxybutyrate (βHB) salt infusion lowers blood glucose in both animals [3][4][5] and humans. 6 Additionally, the consumption of ketone ester (KE) 6 or medium chain triglyceride oils 5 lowers blood glucose, suggesting that ketosis itself has a hypoglycaemic effect that is independent of carbohydrate restriction.
For reasons that are unknown, the glucose-lowering effect of ketosis appears to be greater in patients with type 2 diabetes (T2D) than in healthy controls, 7 but both insulin dependent 8,9 and independent 10 mechanisms have been proposed. As some patients find carbohydrate restriction difficult leading to low long-term adherence 11 and as exogenous ketosis can independently improve glycaemic control (even after a carbohydrate challenge 12 ), there exists a clinical niche for a safe exogenous ketogenic intervention in T2D.
The ketone monoester (KE), (R)-3-hydroxybutyl-(R)-3-hydroxybutyrate, is a drink that, after ingestion, is cleaved by gut esterases into equimolar βHB and (R)-1,3-butanediol. Both molecules enter the portal circulation, and the latter is converted in the liver into βHB. 13 Thus, each equivalent of KE yields two equivalents of βHB.
The pharmacology and safety of KE has been thoroughly studied in animals 14 and in healthy human adults. 15 However, the safety and tolerability of this supplement has not been investigated in patients with T2D. Given the metabolic and physiological peculiarities of this disease, and the potential benefits of exogenous ketosis on glycaemic control, we aimed to study the safety and tolerability of the KE in patients with T2D.
The KE is available as a sports supplement 16 and carries benefits over other ketogenic supplements. First, it provides enantiomerically pure d-βHB, which is important because the body primarily uses the d form of βHB and not the l enantiomer. Second, it is salt-free and so does not impose an undue sodium load, particularly when used chronically. Third, KE can raise βHB in blood to fasting-like levels of >3 mM within minutes. 6

| Recruitment and participant characteristics
Potential participants were invited via paper advertisements glycaemic treatment during the study, their participation was discontinued, and they were excluded from the analysis. Also excluded were patients with serum creatinine >1.5 mg/dl, with elevated (3x the normal upper limit) alanine aminotransferase (ALT) or aspartate amino transferase (AST), with scheduled surgical procedures, or who were pregnant, lactating or planning to become pregnant.

| Study design
Each participant was asked to wear an IS-GM (Abbott FreestyleR; Maidenhead, Berks UK ® ) for 6 weeks. Week 1 served as a baseline.
During weeks 2-5, participants were asked to measure out 25 ml (25 g) of the ketone monoester in a measuring cup and drink it either neat or diluted with the zero calorie artificially sweetened drink of their choice, three times daily. Figure S1 illustrates these interventions. Participants were also asked to record their meal timings, sleep duration, physical activity and symptoms (both positive and negative) in a diary. They were instructed to keep their usual diet and physical activities throughout the study and to allow at least 4 h between KE drinks to elevate serum levels of βHB for at least 12 h every day.

| Metabolites and measurements
Tolerability was assessed with self-recorded open adverse effects diaries following each of the KE drinks. If an adverse effect was registered, participants were asked to rank it as 'mild', 'moderate' or 'severe'. Safety was assessed weekly by measuring peak βHB blood levels and acid-base balance. For measuring peak βHB levels, participants measured their blood βHB using Abbot OptiumNeo ® meters immediately before and 30 min following their third KE drink on their least active day of the week.
Weekly anthropometric measurements (body weight and body composition) and fasting blood and urine samples were collected every 7 days after enrolment. Body composition was measured using fourpoint bioelectrical impedance scale (BF508 ® , Omron Electronics Ltd).
Glycaemic control was assessed using (i) blood HbA1c, (ii) blood fructosamine, (iii) mean blood glucose and (iv) blood glucose time in range. Fructosamine was measured using a semi-automated benchtop analyser (ABX Pentra ® , A11A01679). IS-GM's were placed on participants 1 week before starting KE supplementation and changed every 2 weeks throughout the 6-week trial. Target glucose range was 3.9-10 mmol/L (70-180 mg/dl), and time in range was obtained from the Abbot's Libre View ® interface.
For assessing cardiovascular risk, HOMA-IR and 10-year fatal coronary heart disease risk were calculated using open access calculators available at https://www.dtu.ox.ac.uk/riske ngine/ downl oad.php.
Additionally, plasma total cholesterol, ApoB, triglycerides, HDL, Creactive protein, and ALT and AST were measured by the Department of Clinical Biochemistry at the John Radcliffe Hospital, Oxford, UK.

| Sample size rationale
Considering the 1-month duration of the intervention, fructosamine was used for sample size estimation. Other studies report that a 15% decrease in fructosamine is clinically relevant. 17 Assuming a standard deviation of 20% and given statistical power of 0.80 and alpha of 0.05, 16 participants were required for this study. Sample size calculations were performed using G*Power Software version 3.1. 18

| Analysis
Mean and dispersion values for demographic and biochemical variables were calculated. Two-tailed Student's t-tests for paired samples were used for comparing means. Significance was taken at p < .05.
Data are presented as mean ± SD. All calculations were performed using Microsoft Excel ® . There were fewer than 5% missing data points, which were handled by mean substitution.

| Participant characteristics
Twenty-three patients were initially enrolled. One participant withdrew after 1 week without reporting any adverse event. Another participant completed 5 weeks of the study but was excluded from the analysis because she was prescribed glucocorticoids during her last week of the intervention. Therefore, twenty-one participants (14 males and 7 females), aged 45.1 ± 10.8 years, completed the study. Co-morbidities included hypertension (38%) and dyslipidaemia (28%). Common glycaemic therapies included metformin (90.4%), liraglutide (28%) and empagliflozin (19%) ( Table 1).

| Tolerability and adherence
All participants had >90% adherence, defined as consumption of at least 76 of the maximum 84 KE drinks. Thus, the KE drink was welltolerated. Adverse symptoms were self-recorded by participants following each of their drinks and were ranked as 'mild', 'moderate' or 'severe'. Overall occurrence of adverse symptoms is expressed relative to the total number, 1,588, of drinks consumed throughout the entire study. Mild nausea (5/1,588 = 0.3%), mild headache (5/1,588 = 0.3%) and mild gastric discomfort (9/1,588 = 0.5%) were the only reported symptoms.

| Metabolic outcomes and glycaemic control
Blood electrolytes, acid-base status and renal function were normal throughout the study. βHB blood concentrations were recorded by each participant 30 min following a 25 ml KE drink and ranged from 3.1 ± 0.5 mM to 3.8 ± 0.7 mM ( Table 2).
All anthropometric and cardiovascular risk markers were not different at baseline and follow-up (Table 3).

| DISCUSS ION
The primary purpose of this study was to evaluate the safety and tolerability of KE drinks in patients with T2D. Since no moderate or severe symptoms were recorded, the frequency of any given mild symptoms was 0.5% or less, and as no patient experienced symptomatic hypoglycaemia, it is reasonable to conclude that the KE is safe and well-tolerated. Furthermore, the extremely low number of even mild symptoms is congruent with participants' high adherence to the KE drink regimen.
No statistically significant differences were observed in anthropomorphic or cardiovascular risk parameters, other than those related to glycaemic control. This includes no differences in blood pressure, fasting lipid profile, C-reactive protein and body composition, which contrasts with animal studies that report changes in such metrics. 19 However, there was a small reduction in total weight of 1.3 kg which could have be greater with a longer-term intervention, meriting further research. It is worth highlighting that these metrics were followed primary for safety purposes and that this study was not powered, nor intended to look for, changes in these markers. In other words, it is possible that improvements in weight and markers other than those related to glycaemic control may have been observed had the intervention lasted longer than 1 month or had a larger cohort of patients been studied.
Importantly, all four markers of glycaemic control-HbA1c,

| CON CLUS IONS
These results confirm the safety and tolerability of three times daily KE drinks in patients with T2D. The KE drinks improved four markers of glycaemic control: HbA1c, fructosamine, mean glucose and time in range. Randomized clinical trials are needed to evaluate their clinical efficacy.

CO N FLI C T O F I NTE R E S T
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: The intellectual property covering the uses of ketone bodies and

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.