Barely, and not in a way you can rely on. In a double-blind randomised crossover trial, people who were given coffee, caffeine, decaffeinated coffee or placebo after breakfast went on to eat statistically indistinguishable amounts at a free-access meal four and a half hours later. Their appetite ratings, gastric emptying and plasma glucose did not separate either.
A separate review of the wider literature suggests the one thing that does matter is timing, and that the window where coffee helps is much narrower than the habit assumes.
The trial that tested all four conditions
A 2014 study in Appetite used a double-blind randomised crossover design, meaning the same 12 participants, 9 of them women, went through every condition rather than being split into groups. Average age was 26.3 years and average BMI 22.7.
Each trial started with a standardised breakfast labelled with carbon-13 octanoic acid so gastric emptying could be tracked from breath samples. Participants then had 225 mL of the treatment beverage plus a capsule, and a second identical serving two hours later. Four and a half hours after breakfast they were given free access to a meal, and what they ate was measured.
What came out of it
Energy intake at the test meal was 2118 kJ on placebo, 2128 on decaffeinated coffee, 2287 on caffeine and 2016 on caffeine with decaf. None of the differences were significant. Note the direction as well as the size: the caffeine-only condition produced the highest intake of the four, which is the opposite of what the appetite-suppression story predicts.
Appetite sensations showed nothing beyond a main effect of time, which simply means everyone got hungrier as the morning went on. Plasma glucose showed no treatment differences. Gastric emptying, the mechanism most often invoked for why coffee might blunt hunger, was not significantly different across trials.
Four ways of measuring the same claim, four null results, in a design that controls for individual differences better than a parallel-group study of the same size would.
The size of the study is the obvious objection, and it deserves a straight answer rather than a dismissal. Twelve people is small, and a small trial can miss a real effect. What makes this one more informative than its headcount suggests is the crossover structure: each person serves as their own control across all four conditions, which removes the between-person variation that usually swamps appetite research. It is still a single study in young, healthy-weight adults, and the authors close by saying that caffeine and coffee over longer periods warrant further investigation. They are not claiming the question is shut.
The finding that rescues part of the belief
A 2017 review in the International Journal of Food Sciences and Nutrition, from overlapping authors, gathered the studies on caffeine, coffee, energy intake, gastric emptying, appetite hormones and perceived appetite.
Its conclusion splits on timing. Coffee taken 3 to 4.5 hours before a meal had minimal influence on food and macronutrient intake. Caffeine taken 0.5 to 4 hours before a meal may suppress acute energy intake. The evidence on gastric emptying, appetite hormones and appetite perceptions was described as equivocal.
So the 2014 trial's four and a half hour gap sits at the far end of the window where the review says least happens, which makes the two results consistent rather than contradictory. The practical version is narrow: if caffeine does anything to your intake, it does it close to the meal, not as a morning-long hunger shield. That is a much weaker claim than the one built into skipping breakfast for coffee, and it sits alongside what meal frequency actually changes, which is also less than the habit implies.
What nobody has measured yet
The review is unusually direct about its own gaps. The influence of genetics of caffeine metabolism is unknown, as is the influence of bitter taste phenotype, and it calls for longer controlled studies.
Both of those are real sources of person-to-person variation. People clear caffeine at very different rates, and how bitter a given coffee tastes to you is partly genetic. If either turns out to matter, the population-level null could be hiding two groups cancelling each other out. That is a plausible story and, at the moment, only a story.
Why OffRamp ignores a half point
OffRamp's daily check-in records hunger and energy on a simple 1 to 5 scale. The insights engine that reads those check-ins will not tell you a trend has changed unless it moves by at least half a point week over week, and it will not comment on an energy difference between strength days and other days unless the gap reaches 0.4 points.
Those thresholds were set deliberately high, and this topic is a clean example of why. A morning coffee, a slightly later lunch, a bad night's sleep and a stressful Tuesday all move a self-reported hunger rating around by a fraction of a point, and none of them mean anything. A system that reported every wobble would produce a confident explanation every single week, and most of those explanations would be wrong.
The rule is the same one the trial above demonstrates from the other end. Perceived appetite and measured intake come apart constantly, so a sensation is worth recording and not worth acting on until it persists. Telling food noise from hunger is the same distinction at the level of a single afternoon, and getting hunger cues back is what it looks like when the signal genuinely is changing.
None of this is medical advice. If caffeine affects your sleep, your heart rate or a condition you are managing, that belongs with your prescriber rather than with an appetite study.
