Bertoia · PLoS medicine 2015 · Prospective cohort study · n=133,468

Changes in Intake of Fruits and Vegetables and Weight Change in United States Men and Women Followed for Up to 24 Years: Analysis from Three Prospective Cohort Studies.

Cited 402 times in the scientific literature.

Level 3 - non-randomized controlled study

Non-randomized prospective cohort follow-up study

PubMed 26394033 · doi:10.1371/journal.pmed.1001878 · record verified 2026-08-29

What was done

Researchers analyzed data from three prospective cohorts of 133,468 United States men and women followed from 1986 to 2010 (up to 24 years). They assessed the relationship between changes in the intake of specific fruits and vegetables and weight change across multiple 4-year intervals. Analyses adjusted for changes in other lifestyle factors (including other dietary aspects, smoking status, and physical activity), and results were combined using random-effects meta-analysis.

What was found

Increased fruit intake was inversely associated with 4-year weight change: total fruits -0.53 lb per daily serving (95% CI -0.61, -0.44), berries -1.11 lb (95% CI -1.45, -0.78), and apples/pears -1.24 lb (95% CI -1.62, -0.86). Vegetable intake associations varied: total vegetables -0.25 lb per daily serving (95% CI -0.35, -0.14), tofu/soy -2.47 lb (95% CI -3.09, -1.85), and cauliflower -1.37 lb (95% CI -2.27, -0.47). Conversely, increased intake of starchy vegetables (corn, peas, potatoes) was associated with weight gain (numerical effect sizes not reported in abstract). Vegetables with higher fiber and lower glycemic load had stronger inverse associations with weight change than lower-fiber, higher-glycemic-load vegetables (p < 0.0001).

Why it matters

Broad dietary advice often groups all produce together, but these findings show that high-fiber, low-glycemic fruits and vegetables support weight management, whereas starchy vegetables are linked to weight gain.

Limits

The study is observational, leaving open the potential for residual confounding despite multivariable adjustments. Dietary data were self-reported via food frequency questionnaires, carrying inevitable measurement error. Numerical estimates and confidence intervals for starchy vegetables were omitted from the abstract.

Cited by