Answers are grounded in this paper and cite their sources.
β-Hydroxybutyrate (BHB) is the primary circulating ketone body during nutritional ketosis. Circadian regulation of ketogenesis has been characterized in rodents, but time-of-day variation in blood BHB has not been quantified in free-living humans, and the concurrent relationship between BHB and glucose has not been described at scale. Because people and researchers almost always measure BHB in the morning, alongside fasting glucose, any daily variation would be systematically invisible. To characterize the 24-hour rhythm of blood BHB and its concurrent relationship with glucose in free-living individuals who monitor ketosis, and to estimate how far morning-only sampling misrepresents daily BHB exposure. Retrospective observational study of capillary blood BHB and glucose readings recorded with commercial meters (Keto-Mojo GK+, GKI, and TD-4279) and synchronized to a cloud platform (MyMojoHealth) from January 2018 to August 2025. Feeding, activity, sleep, and light exposure were not controlled or recorded. Rhythms were assessed by single-component cosinor analysis and linear mixed-effects models with random intercepts for user; the BHB-glucose relationship was evaluated by linear and logistic regression. The circadian analysis included 140,024 users (11,858,248 BHB readings); the concurrent relationship analysis included 138,330 users (11,783,637 paired BHB-glucose readings). Self-reported sex was the only participant characteristic available, in 8% of users. BHB showed a significant 24-hour rhythm (amplitude 0.085 mmol/L, 7.42% of the mesor; acrophase 19:15; p < 0.001), with evening levels 15.14% higher than morning (1.23 vs. 1.06 mmol/L, p <2e-16). The conventional morning testing window coincided with the trough of the cycle. Males showed higher BHB and a larger morning to evening contrast than females ( p <2e-16). Each 1 mmol/L increase in BHB was associated with a 6.19 mg/dL lower glucose (95% CI 6.09 to 6.29; p < 0.001), explaining 9.0% of glucose variance. The proportion of readings in the 70 to 100 mg/dL range rose from 48.3% at BHB 0.1 to 0.5 mmol/L to 75.5% at BHB 2.0 to 3.0 mmol/L. BHB and glucose followed opposing 24-hour profiles. In sensitivity analysis, an unmeasured confounder would need to be associated with both BHB and glucose by a risk ratio of at least 1.97 per 1 mmol/L, and 2.89 per 2 mmol/L, to explain away the association (E-values). Blood BHB follows a consistent 24-hour pattern with a morning trough and an evening peak among individuals monitoring ketosis, and higher BHB is accompanied by lower concurrent glucose across the nutritional ketosis range. Clinically, morning-only measurement samples the lowest point of the cycle and underestimates the daily average by approximately 15%, so clock time should be recorded and standardized whenever BHB is used to judge dietary adherence, to titrate ketogenic therapy, or to compare studies. Because feeding, activity, and sleep were uncontrolled, the pattern described here is the combined output of endogenous timing and daily behavior, and it should not be extrapolated to the general population or to insulin-deficient diabetes.
No takes yet. Share an insight, caveat, or question.
Soto-Mota et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: