The present study demonstrates for the first time that KME ingestion prior to a meal can robustly reduce PBG in individuals with type 2 diabetes. There was a 28% reduction in the rate of glucose appearance over the first 2 h following meal ingestion, which resulted in a significant 18% reduction in PBG in the first 2 h and an overall 12% reduction in glucose AUC across the entire 4 h. Contrary to our hypothesis, EGP was unaffected by KME ingestion and played no significant role in reducing postprandial glucose concentrations. The postprandial glucose-lowering effect seen here was greater than that observed in individuals with normal [3, 4] or impaired glucose tolerance [5], indicating that the mechanism(s) of action for pre-meal KME-mediated glucose reduction appear to be intact in individuals with type 2 diabetes.

The mechanism(s) by which KME reduced PBG fits with a delay in, rather than complete inhibition of, glucose absorption, as total glucose appearance across the entire 4 h postprandial period was unaltered due to a greater RaT in the later postprandial period compared with the control condition. The delay could be due to slowed gastric emptying; however, the effect of KME ingestion on gastric emptying remains inconclusive [6, 7]. Similarly, it has been repeatedly shown that meal preloads increase the secretion of gut peptides, influencing pancreatic hormone secretion and modulating gastrointestinal function [6, 7]. As such, further research on the effects of KME ingestion on gastrointestinal transit and function (including hormonal secretion and gastric emptying) is warranted. In this regard, it should be noted that the placebo in our study contained no caloric value; future work should compare KME with other energy-matched meal preloads. Additionally, bitter taste-sensing type 2 receptors (TAS2Rs) are thought to play a role in regulating PBG via upregulated glucagon-like peptide-1 signalling [8] and it is possible that the bitterant in the control condition may have decreased PBG and dampened the differences between conditions. Furthermore, the potential for metformin or other glucose-lowering medication to interact with the effect of KME on reducing, or delaying, PBG warrants further investigation.

We also report a potent reduction in circulating NEFA concentrations after KME ingestion compared with CON ingestion in individuals with type 2 diabetes, in the absence of a difference in circulating insulin levels between conditions. While insulin concentrations did not differ between conditions, this is not necessarily indicative of identical insulin secretion or clearance in both conditions. The observed reduction in NEFA levels fits with a role of β-OHB as an endogenous ligand for the hydroxycarboxylic acid receptor 2 (HCAR2) [9], which inhibits adipose tissue lipolysis; however, we did not employ a tracer to directly measure the rate of lipolysis, which has been performed previously in healthy individuals [10]. Pharmacologically manipulating circulating NEFA levels in humans can rapidly modulate insulin-stimulated glucose disposal in skeletal muscle [10] and EGP in the liver [11], so it is surprising that KME ingestion reduced Rd but did not measurably reduce EGP. This is particularly notable given previous observations of reduced EGP when circulating β-OHB was increased via oral ketone administration in healthy individuals [12, 13]. The lack of suppression of EGP with KME in our study could be due to measurement error, which could be reduced by using more advanced computational methods with variable tracer infusions to mimic the expected effect. However, approaches similar to ours have previously yielded accurate results [14] and detected differences in EGP with pharmacological intervention in individuals with type 2 diabetes [15].

We acknowledge several other caveats when interpreting our results. First, the postprandial period we investigated was in the context of participants breaking their fast and it is unclear whether the PBG-lowering effect would be retained for subsequent meals and how it might be modulated by circadian fluctuations in metabolism. Additionally, the observed lowering of PBG, although marked, was relatively short-lived, occurring only in the initial postprandial period. Finally, although we included both sexes in our study, we were not powered in this proof-of-concept trial to determine if (and to what extent) biological sex has the potential to moderate the observed effect, which should be considered in future investigations.

In conclusion, the ingestion of a KME preload prior to a MMTT led to a robust decrease in PBG in individuals with type 2 diabetes. Clinical trials to assess the efficacy of daily KME ingestion as a treatment option for type 2 diabetes are warranted.