Dong · PloS one 2020 · meta-analysis of randomized trials · n=12 studies

The effects of low-carbohydrate diets on cardiovascular risk factors: A meta-analysis.

Cited 109 times in the scientific literature.

Level 1 - systematic review of randomized trials

Meta-analysis of randomized controlled trials

PubMed 31935216 · doi:10.1371/journal.pone.0225348 · record verified 2026-08-30

What was done

Four electronic databases (PubMed, Embase, Medline, Cochrane Library) were searched through November 2018 for randomized controlled trials evaluating low-carbohydrate diets. Twelve randomized trials measuring body weight, blood pressure, total cholesterol, HDL-C, LDL-C, triglycerides, and fasting blood glucose were included and analyzed using fixed- or random-effects models, with stratification by intervention duration.

What was found

Compared to control diets, low-carbohydrate diets showed: - Triglycerides: -0.15 mmol/L (95% CI -0.23 to -0.07); -0.23 mmol/L (95% CI -0.32 to -0.15) for <6 months, and -0.17 mmol/L (95% CI -0.32 to -0.01) for 12–23 months. - Body weight: -1.58 kg (95% CI -1.58 to -0.75); -1.14 kg (95% CI -1.65 to -0.63) for <6 months, and -1.73 kg (95% CI -2.70 to -0.76) for 6–11 months. - Systolic blood pressure: -1.41 mmHg (95% CI -2.26 to -0.56). - Diastolic blood pressure: -1.71 mmHg (95% CI -2.36 to -1.06). - HDL-C: increase of 0.1 (unit erroneously reported in abstract as mmHg; 95% CI 0.08 to 0.12). - Total cholesterol: +0.13 mmol/L (95% CI 0.08 to 0.19). - LDL-C: +0.11 mmol/L (95% CI 0.02 to 0.19). - Fasting blood glucose: +0.03 mmol/L (95% CI -0.05 to 0.12, not significant).

Why it matters

Low-carbohydrate diets produce small favorable improvements in blood pressure, weight, and triglycerides, counterbalanced by modest increases in total and LDL cholesterol.

Limits

The total number of participants across the 12 included studies is not stated in the abstract. The abstract includes obvious typographical errors in reporting (HDL-C unit given as mmHg; lower 95% CI bound for overall weight loss identical to the point estimate). Specific carbohydrate thresholds, comparator diets, dietary adherence, and clinical cardiovascular events were not reported.

Cited by