The long-run outcome data behind “diets don’t work” — and the argument attached to it, kept separate.
What the evidence says
- Diets reliably produce short-term loss of 5–10% of starting weight, and the losses are not maintained. Mann and colleagues quote an earlier review (Garner & Wooley, 1991): ”It is only the rate of weight regain, not the fact of weight regain, that appears open to debate”[1]
- In 14 observational studies with four-plus-year follow-up, average initial loss was 14 kg and participants regained all but 3 kg[1]
- Across 8 studies reporting counterproductive outcomes, an average of 41% of participants (range 29–64%) weighed more at follow-up than before they dieted[1]
- In randomised trials, dieters maintained an average loss of only 1.1 kg. The authors’ own most generous reading is that dieting slows the slight weight gain that comes with age[1]
- *Of 10 prospective non-randomised studies, 7 found that dieting predicted weight gain over time, and in one of them, of over 19,000 healthy older men, ”One of the two best predictors of weight gain over the four years was having lost weight on a diet at some point during the years prior to baseline”*[1]
- Mann’s central methodological claim is that the literature is biased toward showing success, on four counts: follow-up rates averaging 33% (8 of 14 studies below 50%), self-reported weights (56% of participants; a separate meta-analysis of 15 studies, which Mann cites, found people underestimate their weight by about 2.1 kg), diet confounded with exercise, and participants entering further diets between study end and follow-up[1]
- The review found no consistent evidence that dieting produces health improvements independent of weight change; where benefits appeared, they could not be separated from exercise, medication or other changes[1]
- Weight cycling, in observational studies the review cites, is linked to increased all-cause mortality and cardiovascular mortality, and associated with myocardial infarction, stroke, diabetes, raised blood pressure and suppressed immune function. Whether intentional dieting is to blame is argued in the review, which concludes that ”the balance of evidence does seem to implicate intentional weight loss in adverse health outcomes”[1]
- In one comparison of diet-only, exercise-only and combined interventions (Skender 1996), the diet-only group ended above baseline while ”the groups that included exercise remained 2.5 kg (5.5 lb) below baseline”. Separately, of a lifestyle intervention the review says ”participants in the lifestyle intervention engaged in large amounts of physical activity (averaging 227 minutes per week), and this may be the potent factor”[1]