Afternoon Fasting Beats Morning and Evening for Blood Sugar Control, New Study Finds

When you start a 24-hour fast turns out to matter almost as much as the fast itself. A new randomized crossover study published in Frontiers in Nutrition found that adults with overweight or obesity who began fasting at 1:00 pm had lower average glucose levels, a smaller post-meal glucose spike, and deeper ketosis at the end of the fast compared with those who started fasting at 8:00 am or 6:00 pm. The same 24 hours, three different clocks, three different metabolic results.

Afternoon fast start time lowers mean glucose and post-meal spike

Twenty-four adults completed all three fasting conditions in randomized order, each followed by a 75-gram oral glucose tolerance test (OGTT). Continuous glucose monitors tracked interstitial glucose throughout each 24-hour window and the two-hour OGTT. The afternoon condition produced the lowest mean 24-hour glucose and the smallest post-prandial glucose area under the curve (AUC) following the standardized meal at the end of the fast. Morning and evening fasts did not differ significantly from each other on either of those two metrics. The gap is not explained by differences in insulin or glucagon: plasma levels of both hormones were measured at the start and end of each fast and 30 minutes after the OGTT, and none of those measurements differed significantly across the three conditions. Something downstream of those classic regulators appears to be doing the work.

The circadian system's tight control of glucose metabolism is well established in the literature. The body's molecular clocks, active in the liver, pancreas, and muscle, set the pace for insulin secretion, glucose uptake, and glycogen storage across the 24-hour day. What this study adds is a direct experimental comparison of how the timing of a full-day fast interacts with that rhythm.

Ketosis depth differs by fasting start time

Beta-hydroxybutyrate (BHB), a ketone body used as a proxy for metabolic switching, was measured at the end of each fast. The threshold the researchers used for clinically meaningful ketosis was 0.5 mmol/L. Participants who fasted from 1:00 pm and those who fasted from 6:00 pm both crossed that threshold by the end of their respective fasts. Participants who fasted from 8:00 am did not. That result is counterintuitive at first glance: a morning fast extends overnight fasting and might be expected to push ketone production furthest. The data suggest that the circadian phase at which the fast ends, not just how long glucose has been withheld, shapes how aggressively the liver shifts toward fat oxidation.

Research on chrononutrition and metabolic outcomes has increasingly pointed in this direction: aligning eating and fasting windows with the body's internal clock, rather than simply restricting calories or compressing eating windows, appears to extract additional metabolic benefit. Harvard's nutrition team has similarly noted that earlier eating windows in intermittent fasting are associated with better blood sugar regulation, though the mechanism has remained murky.

Randomized crossover design strengthens causal inference

The study's crossover structure is its main methodological strength. Each participant served as their own control across all three fasting conditions, eliminating between-person variation in baseline metabolism, body composition, and habitual diet. The 75-gram OGTT is a standardized clinical tool for assessing glucose tolerance, and the continuous glucose monitors provided granular, minute-by-minute glucose trajectories rather than isolated blood draws. Plasma insulin and glucagon sampling at three standardized time points per condition added hormonal context.

The sample size, 24 participants, is modest, and CGM data was missing for 18 conditions (out of a total of 72, given three conditions per participant) because some monitors were removed prematurely. That data loss introduces some uncertainty into the glucose trajectory analysis, though the researchers accounted for it in their statistical models. The field of chrononutrition as a whole still contends with scarce, sometimes conflicting clinical findings and a tendency toward small, short-duration trials, a limitation this study shares.

Morning fast produces better glucose tolerance on refeeding

One finding runs against the afternoon-is-best narrative and deserves attention. The glucose AUC above baseline at the conclusion of the fast, which reflects how the body handles the OGTT immediately after breaking the fast, was optimal after the morning fast, not the afternoon fast. That means participants who fasted from 8:00 am showed the most favorable acute glucose tolerance when they ate again, even though their mean in-fast glucose and ketosis depth were worse than in the afternoon condition.

The researchers note this conflicts with most existing literature on morning versus afternoon glucose responses, where afternoon tolerance is typically inferior. One possible explanation is that the physiological preparation for feeding, priming of insulin secretion and hepatic glucose uptake, follows its own circadian trajectory that is distinct from the mechanisms governing in-fast glucose and ketone dynamics. The finding also echoes a broader debate in time-restricted eating research: a Science Translational Medicine analysis raised questions about whether the metabolic benefits of compressed eating windows are driven by circadian alignment at all, or whether unintended calorie restriction deserves most of the credit.

What the findings mean for fasting protocol design

For the roughly 40% of American adults who meet criteria for overweight or obesity and who are considering fasting as a metabolic intervention, these results offer a specific and testable refinement: starting a 24-hour fast around lunchtime rather than at breakfast may lower average blood glucose and push the body further into ketosis by the time the fast ends. The effect appears independent of the hormonal signals most clinicians reach for first when explaining glycemic change.

What the study cannot yet answer is whether these advantages persist over weeks or months of repeated afternoon fasting, whether they generalize to people with normal weight or to those with diagnosed type 2 diabetes, and whether the morning refeeding advantage represents a genuine clinical benefit or a transient artifact of measurement timing. Larger trials with longer follow-up and tighter CGM compliance are the clear next step. For now, the data make a credible case that circadian biology should be a design variable in fasting protocols, not an afterthought.