No, on the randomised evidence. Pooling 129 short-term comparisons, people ate less after a low-energy sweetener than after sugar and no more than after water. In sustained trials running from 4 weeks to 40 months, replacing sugar with a sweetener came with roughly 1.35 kg less body weight, not more.
The studies that made sweeteners look fattening are a different kind of study, and the difference explains almost the whole argument.
What the pooled trials found
A 2016 systematic review in the International Journal of Obesity set out to test the null hypothesis that low-energy sweeteners have no effect on energy intake or body weight, and gathered animal studies, prospective cohorts and randomised trials rather than picking one type.
In short-term randomised trials, 129 comparisons, total energy intake was about 94 kcal lower when people had a low-energy sweetener before an ad libitum meal instead of a sugar-sweetened food or drink. Against water the difference was about 2 kcal, with a confidence interval straddling zero. In sustained trials, nine comparisons against sugar gave about 1.35 kg less body weight, and three comparisons against water gave about 1.24 kg less.
Why the observational studies say the opposite
The same review pooled 12 prospective cohort studies in humans and found the association between sweetener use and BMI was essentially nothing, at -0.002 kg per square metre per year, with a confidence interval crossing zero in both directions. Individual cohorts have reported positive associations, and those are the studies that generate the headlines.
The authors name the likely reason directly: reverse causation. People do not take up diet drinks at random. They take them up because they are already gaining weight, already have a higher BMI, or have already been told to cut sugar. A cohort study then records sweetener use and later weight gain and finds them travelling together, because the weight concern came first.
This is the single most useful thing to understand about the whole sweetener argument. Both literatures are real. Only one of them randomises who gets what, and that is the one that can answer a causal question.
It is worth being clear about what the review did and did not claim, since its own wording is more careful than most of its coverage. It says the preponderance of evidence from all human randomised trials indicates sweeteners do not increase energy intake or body weight, whether the comparison is a caloric control or a non-caloric one such as water. Its stronger sentence, that using sweeteners in place of sugar leads to reduced intake and weight, is offered as the balance of evidence rather than as a settled result, and the comparison against water carries a hedging "possibly". That is a review reporting a direction confidently and a magnitude cautiously.
The animal studies, and why they are weaker than they look
The review also covered 90 animal studies, and in 62 of them sweetener exposure either did not affect body weight or decreased it. Of the 28 that reported an increase, 19 used a specific learning paradigm comparing a sweetener against glucose exposure.
The authors' own assessment is that most animal studies did not mimic how humans actually consume these compounds. That matters because the popular theory, that a sweet taste without energy trains the body to mishandle real sugar later, comes largely out of exactly those paradigms. The theory is not silly. It just has not reproduced in people eating normally.
Where a real effect does show up
Weight is not the only outcome worth measuring, and on a different outcome the picture changes. A 2022 randomised controlled trial published in Cell gave 120 healthy adults sachets of saccharin, sucralose, aspartame or stevia for two weeks, at amounts below the accepted daily intake, against controls receiving the sachet vehicle or nothing.
Every one of the four altered the stool and oral microbiome and the plasma metabolome. Saccharin and sucralose significantly impaired glycemic responses. The effects were personalised rather than uniform, which is why the paper's title says microbiome-driven.
So the honest summary has two halves that point in different directions. On body weight, the randomised evidence does not support the backfire story. On the microbiome and short-term glucose handling, there is a measured effect in humans. Anyone giving you one half without the other is arguing rather than reporting. The gut side of that connects to what the microbiome does around weight regain, which has the same mouse-versus-human caveat running through it.
Why OffRamp has no position on your coffee
OffRamp tracks two things. Grams of protein against a floor set at 1.6 g per kilogram of your goal weight, and a 7 day rolling weight average compared against your baseline. There is no ingredient scoring, no food grading, no red or green label on anything you log.
That is a real decision and this topic is a good illustration of why it was made. A food-scoring system built in 2016 would have marked sweeteners green on the trial evidence. One built from the cohort headlines would have marked them red. The evidence has not actually changed much in that window, but the popular reading of it has swung twice, and an app that encoded either reading would now be confidently wrong at users who did nothing but keep logging.
What the log does instead is count. If swapping sugar for a sweetener helps you hold your pattern, the trend line will show it. If a diet drink is standing in for a meal and your protein floor is quietly going unmet, the hit rate will show that too. Sugar cravings and quieting food noise cover the part of this that is actually about wanting the sweet thing rather than about the chemistry.
None of this is medical advice or guidance about any medication. If you have diabetes or another condition where glucose control matters, the 2022 glycemic finding is a conversation for your prescriber and not a reason to change anything on your own.
