Inter-meal interval and hunger
The question
How long a gap between meals is physiologically necessary for gastric emptying and satiety signal recovery, and what minimum spacing can a meal schedule reasonably use?
What the evidence says
Gastric emptying sets a floor, and it depends on what’s in the meal. The most direct data comes from Farivar et al. (2026), who used MRI to track gastric volume every two hours after standardized high-fat, high-carb, and high-protein meals in 15 healthy adults. Half-emptying time (the point at which half the meal has left the stomach) was 132±33 minutes for the high-carb meal, 158±37 minutes for the high-protein meal, and 188±41 minutes for the high-fat meal. Full emptying takes roughly twice the half-time, so a typical mixed meal is mostly, but not completely, out of the stomach by 3-4 hours. This lines up with older gastric-emptying literature (Camilleri’s group has published extensively on measurement methods and physiology, e.g. Szarka & Camilleri 2009): fat reliably slows emptying, protein slows it less, and liquids and simple carbs clear fastest.
The practical read: eating again well before the prior meal has substantially emptied doesn’t cause harm, but it does mean the new meal is digesting on top of unfinished business from the last one, which muddies both the glycemic response and the felt hunger signal.
Satiety hormones track digestion, not a fixed clock. GLP-1, PYY, and CCK are released from the gut in response to food contact and nutrient absorption, and their appetite-suppressing effect roughly follows the emptying curve rather than a separate timer. D’Alessio’s 2008 review lays out the mechanism: all three peptides rise within the first hour after eating and decline over the following 2-4 hours as the meal clears, with fat and protein prolonging the response more than carbohydrate. GLP-1 itself is degraded within minutes (its circulating half-life is on the order of 2-3 minutes, per Maselli & Camilleri’s review of GLP-1 physiology), but it’s continuously resecreted as long as nutrients remain in the gut, so the appetite effect lasts far longer than the molecule does. Ghrelin, the hunger-signaling hormone, works in reverse: it’s suppressed after eating and rises again as the stomach empties and nutrient absorption tapers off. None of this converges on a sharp cutoff. It’s a gradient, not a threshold.
Protein slows the whole cycle down. Leidy et al. (2011) randomized overweight men to higher- or normal-protein diets, each tested at both 3 meals/day (every 5h) and 6 meals/day (every 2h). The higher-protein diet produced meaningfully more fullness and lower late-night hunger regardless of meal frequency, and within the higher-protein group, 3 meals/day produced better evening fullness than 6 meals/day. That’s a direct signal that spacing protein-rich meals further apart doesn’t cost satiety, and may help it, echoing the gastric-emptying data showing protein and fat extend the digestive window.
Does eating more often actually help hunger control? The evidence here is thinner than the physiology suggests it should be. Speechly & Buffenstein (1999) found that lean men eating more frequently reported better appetite control and ate less at a subsequent test meal, a small but often-cited result in favor of more frequent eating. But the largest synthesis to date points the other way. Schwingshackl et al. (2020), a systematic review and network meta-analysis of 22 RCTs (n=647) comparing isocaloric meal frequencies from 1 to 8+ meals/day, found that energy intake was not affected by meal frequency, and if anything, fewer meals per day (1-2) modestly outperformed more (3-6) for body weight and waist circumference, though certainty was moderate at best and the mechanism isn’t satiety-specific. Together these results say meal frequency by itself is not a lever Eat On Pace should pull for appetite control. The gap between meals matters for comfort and digestion, not because more frequent eating suppresses hunger better.
No dedicated trial tests the minimum comfortable gap directly. Nobody has run the study Eat On Pace actually wants: randomize adults to different fixed inter-meal intervals and measure hunger, fullness, and adherence. What exists is physiology (gastric emptying, hormone kinetics) that bounds the plausible range, plus indirect frequency studies that rule out “more meals = better satiety” as a general rule. The 2-4 hour window this doc lands on is inferred from the emptying and hormone data, not measured directly as a satiety-recovery interval.
Key studies
| Study | Year | Design | n | Finding |
|---|---|---|---|---|
| Farivar et al. | 2026 | MRI gastric-emptying study | 15 healthy adults | Half-emptying time: 132±33 min (high-carb), 158±37 min (high-protein), 188±41 min (high-fat) |
| D’Alessio | 2008 | Narrative review | N/A | CCK, GLP-1, PYY rise within ~1h post-meal and decline over 2-4h as gastric contents clear; mechanism review, not a trial |
| Leidy et al. | 2011 | RCT (12 weeks) | 27 overweight/obese men | Higher-protein diet improved fullness regardless of meal frequency (3 vs. 6 meals/day); 3 meals/day gave better evening fullness than 6 within the high-protein group |
| Speechly & Buffenstein | 1999 | RCT (acute, crossover) | 6 lean men | More frequent eating associated with better appetite control and lower subsequent intake; small sample |
| Schwingshackl et al. | 2020 | Systematic review + network meta-analysis (RCTs) | 22 RCTs, 647 participants | Isocaloric meal frequency (1-8+ meals/day) has no effect on energy intake; fewer meals/day (1-2) modestly outperformed 3-6 for body weight and waist circumference |
DOIs: 10.1007/s00414-026-03836-8 | 10.1177/0148607108322401 | 10.1038/oby.2010.203 | 10.1006/appe.1999.0265 | 10.1093/advances/nmaa056
Confidence rating and why
Emerging.
The gastric-emptying physiology (Farivar 2026, and the broader Camilleri-adjacent literature it sits in) is solid, controlled, mechanistic data, but it measures stomach contents, not hunger or adherence. The hormone review (D’Alessio) explains mechanism well but isn’t a quantitative trial. Leidy 2011 is a real RCT and directly relevant, but it’s one study in one population (overweight men) testing two specific frequencies, not a dose-response curve across intervals. Schwingshackl 2020 is the strongest piece of evidence in this doc by design (systematic review, 22 RCTs) but it answers a different question: whether frequency changes body weight and intake, not what the minimum comfortable gap is. No study directly manipulates inter-meal spacing and measures hunger recovery as the outcome.
Because the minimum-gap number below is inferred by triangulating emptying rates and hormone kinetics rather than read off a trial that measured the thing Eat On Pace needs, this rates Emerging rather than Moderate. The direction (mixed meals need roughly 2-4 hours before hunger and digestion meaningfully reset) is defensible and consistent across sources; the exact hour is a reasonable estimate, not a measured value.
What Eat On Pace does with this
Optimizer constraint: hard minimum inter-meal gap.
Eat On Pace enforces this as a hard floor in the timing solver rather than a soft penalty, because placing meals closer than the stomach can plausibly handle produces a schedule that’s not just suboptimal but physiologically incoherent (a second meal landing on top of an unemptied first one).
Default floor: 2.5 hours between any two meal slots, regardless of composition. This sits below the fastest-clearing case in Farivar 2026 (high-carb half-emptying at ~2.2h) so it doesn’t block genuinely light or carb-forward meals, while still ruling out back-to-back slots that no gastric-emptying data supports.
Composition-aware adjustment (applied when meal macros are known at scheduling time):
- Meals flagged high-protein or high-fat (roughly matching the protein-per-meal.md floor or a fat-dominant profile): minimum gap rises to 3.5 hours, reflecting the slower emptying Farivar 2026 measured for those meal types.
- Small or liquid-dominant meals (shakes, light snacks under ~300 kcal): floor can relax to 2 hours, since these clear fastest and the Leidy/Speechly data doesn’t show harm from closer spacing when total daily protein and calories are controlled.
This is a hard constraint, not a soft scoring term: meal placements violating the floor are infeasible and get filtered from the optimizer’s candidate space, same treatment as the wake/sleep and meal-count constraints.
How the “Inter-meal intervals” preference level should scale on top of this. The Low/Balanced/High preference in the optimization-preferences panel does not touch the hard floor above; per the panel spec, preference levels only tune soft objective weights, and per-meal physiological floors are never weakened by user preference. What the preference level controls is a soft spacing term layered on top of the hard minimum, pushing gaps wider than the floor when the user says they care more about it:
spacing_score = -k_spacing × max(0, gap_target - actual_gap)- Low: gap_target ≈ the hard floor itself (2.5-3.5h depending on meal composition). The soft term does almost nothing beyond the constraint already enforces; users who set this are telling Eat On Pace they don’t mind tightly packed meals.
- Balanced (default): gap_target ≈ 3.5-4h, giving the optimizer a mild pull toward the middle of the plausible comfort window without fighting other constraints (training windows, protein distribution, wake/sleep).
- High: gap_target ≈ 4.5-5h, approaching the point where a meal is close to fully cleared before the next one starts, useful for users who report feeling sluggish or bloated when meals land close together.
k_spacing should be weighted comparably to other soft dimension terms (caloric-distribution.md’s k_CD is a reasonable reference point), scaled down given this doc’s Emerging rating relative to caloric-distribution.md’s Moderate. Because the underlying evidence doesn’t establish a precise ideal gap, this term should never dominate harder-evidenced terms like the per-meal protein floor or training-adjacency scoring.
Evidence card copy (what users see): “Your stomach needs time to clear a meal before the next one lands well. Mixed meals take roughly 2-4 hours to mostly empty, longer for meals heavy in protein or fat. We use this to set a minimum gap between meals; the level you choose here controls how much extra breathing room the schedule gives beyond that minimum. The evidence behind the exact number is still thin, so we treat it conservatively.”
Features it enables or shapes:
- Hard minimum-gap constraint in the schedule optimizer (composition-aware)
- Soft spacing preference term, tunable via the “Inter-meal intervals” row in the optimization-preferences panel
- Basis for a future “your meals are landing close together” nudge if logged meals repeatedly violate the soft target
What we do NOT claim
- That eating more frequently improves hunger control or metabolic outcomes in general. Schwingshackl 2020, the strongest evidence in this doc, found no effect of meal frequency on energy intake and a slight edge for fewer meals on weight and waist circumference. Eat On Pace does not use inter-meal spacing as a lever for “eat more often to control hunger.”
- That there’s a precise, measured minimum gap below which hunger control fails. The comfortable spacing we work from is triangulated from gastric-emptying and hormone-kinetics data, not read off a trial that directly tested spacing against hunger outcomes. The exact spacing is a modelling choice, not a number the evidence pins down.
- That the spacing we use is a hunger or metabolism guarantee. It reflects digestive comfort and subjective fullness, not a proven effect on appetite control or metabolic outcomes.
- That any adjustment to the gap based on meal composition is a precise threshold. Where we account for heavier meals clearing more slowly, that’s a conservative estimate from one MRI study (Farivar 2026, n=15) and general gastric physiology, not a dose-response curve across meal types.