What Vegan Diet Trials Actually Show for Heart Health

The numbers are useful—and smaller than the slogans

A vegan diet can improve several cardiometabolic risk markers within a few months. The best randomized-trial synthesis estimates an average LDL cholesterol reduction of 10.4 mg/dL, weight loss of 2.52 kg, and HbA1c reduction of 0.37 percentage points versus omnivorous or mixed-diet controls. Those are worthwhile changes. They do not show diabetes reversal, prove fewer heart attacks, or establish that vegan eating beats every well-designed Mediterranean or medical nutrition plan.

The most data-dense analysis is Termannsen et al.'s 2022 meta-analysis of 12 randomized controlled trials involving 609 participants. Its median follow-up was 16 weeks. That duration can show whether blood lipids, glucose, and weight move; it cannot answer whether people maintain the diet or the weight loss for years.

A plant-predominant diet is a legitimate tool for lowering cardiometabolic risk markers. The trial evidence gives it a measured effect size, not a halo.

What changed in randomized trials

Compared with omnivorous or mixed-diet control patterns, vegan diets produced statistically significant average changes in the Termannsen analysis:

• LDL cholesterol: −0.27 mmol/L, or about −10.4 mg/dL (95% CI, −0.41 to −0.14 mmol/L).
• Total cholesterol: −0.49 mmol/L, or about −18.9 mg/dL (95% CI, −0.65 to −0.33 mmol/L).
• Body weight: −2.52 kg (95% CI, −3.52 to −1.52 kg).
• BMI: −0.94 kg/m² (95% CI, −1.28 to −0.60 kg/m²).
• HbA1c: −0.37 percentage points (95% CI, −0.56 to −0.18).
• Fasting glucose: −0.40 mmol/L (95% CI, −0.62 to −0.18 mmol/L).
• Systolic blood pressure: −2.41 mmHg (95% CI, −4.54 to −0.28 mmHg).

These are average differences, not guarantees for an individual. They also compare vegan diets mainly with mixed or omnivorous patterns, not with optimized DASH, Mediterranean, or diabetes-focused diets. A 10 mg/dL LDL reduction is clinically meaningful as a risk-factor change, but the trials measured LDL rather than cardiovascular events, hospitalizations, or mortality.

The lipid panel did not improve across the board

The same meta-analysis found that HDL cholesterol fell by −0.10 mmol/L (95% CI, −0.15 to −0.05 mmol/L). Triglycerides changed by −0.06 mmol/L, but the 95% CI ranged from −0.22 to 0.11 mmol/L and crossed zero, so the pooled result was not statistically significant. Diastolic blood pressure was also non-significant.

That does not let anyone calculate a net change in cardiovascular events from the lipid results. LDL has a well-established role in atherosclerotic risk, while the meaning of a diet-induced fall in HDL depends on the broader metabolic context. The defensible statement is narrower: vegan diets lowered LDL and total cholesterol in these short trials, lowered HDL as well, and did not produce a statistically significant triglyceride reduction.

Better glucose control is not diabetes reversal

For people with type 2 diabetes, Lv et al.'s 2025 systematic review and meta-analysis found vegetarian and vegan diets reduced HbA1c by roughly 0.3 to 0.4 percentage points versus control diets, with lower fasting glucose as well. That can matter clinically, especially alongside medication adjustment and weight management. It is not evidence that a diet reverses type 2 diabetes.

"Reversal" requires more than an improvement in a group average. Claims usually imply HbA1c below 5.7% and sustained control. The available randomized evidence does not show that outcome. It shows modest improvement in a surrogate marker over the follow-up periods studied.

Mechanistic work offers plausible explanations without settling the clinical question. Dimba et al.'s 2024 review reported that plant-based and Mediterranean diets produced comparable glycemic improvements in prediabetes and type 2 diabetes contexts while examining gut-microbiota-related pathways. Microbiome changes can help explain a hypothesis; they do not prove that one dietary pattern prevents cardiovascular events.

The key comparison may be food quality, not animal-food exclusion

A broader 2022 systematic review by Remde et al. covered 84 interventional trials of plant-predominant eating patterns. It found that these patterns matched or sometimes exceeded American Heart Association, American Diabetes Association, and Mediterranean diets on lipid and glycemic outcomes. It did not establish clear superiority in acceptability or sustainability over other well-implemented health-oriented diets.

The most useful long-term comparison in this evidence base comes from the DIRECT-PLUS randomized trial. Over 18 months, 294 participants assigned to a green Mediterranean diet enriched with Mankai, green tea, and walnuts lost about 6.2 kg. The standard Mediterranean group lost about 5.4 kg, and the healthy-dietary-guidelines group lost about 1.5 kg. The green Mediterranean diet also produced greater reductions in visceral fat and cardiometabolic markers.

This was not a vegan-versus-vegan trial, so it cannot prove that animal-food exclusion lacks value. It does show that adding more plant foods within a flexible Mediterranean framework can produce substantial results. For many people, the practical target may be more whole plants and fewer refined, energy-dense foods—not dietary purity.

Why the weight-loss result needs restraint

Most vegan-diet trials were not calorie-matched. In the reviews, the weight advantage narrowed substantially in calorie-matched analyses. That pattern points toward lower energy density and lower calorie intake as an important explanation for the average 2.52 kg difference. It does not support a claim that vegan diets create a special metabolic advantage independent of energy balance.

Food composition still matters. In a small randomized crossover trial of about 42 participants, Krenek et al. in 2024 found that two whole-food plant-based vegan conditions outperformed an omnivorous pattern for LDL cholesterol and lipoprotein(a), while the lower-extra-virgin-olive-oil version produced greater lipid lowering than the higher-EVOO version. The study suggests that fat quantity and source can change the response within a vegan diet. Its size and crossover design limit what it can say about people with established cardiovascular disease or severe obesity.

What the evidence cannot answer yet

There is no adequately powered randomized trial here for major adverse cardiovascular events, all-cause mortality, or long-term weight maintenance. There are also no adequately powered trials that isolate effects in people with BMI of 35 or higher or established cardiovascular disease. A 2021 review in postmenopausal women found variable and inconsistent cardiometabolic effects across whole-diet interventions, which argues against assuming that pooled results apply equally to every higher-risk group.

Non-randomized comparisons deserve less weight when judging cause and effect. For example, a 2024 cross-sectional comparison of habitual low-carbohydrate, high-fat, omnivorous, vegan, and vegetarian eaters found no uniform metabolic advantage for vegan or vegetarian patterns across measures. People choose diets for many reasons, so this design cannot determine what caused their biomarker differences. It does, however, caution against treating real-world dietary labels as biologically uniform interventions.

A practical verdict

If your current diet is heavy in refined foods and light on plants, moving toward a well-formulated whole-food plant-predominant pattern is likely to improve LDL cholesterol, body weight, and glycemic markers within roughly 12 to 16 weeks. Check the actual numbers: the average changes in the vegan RCT meta-analysis were about 10 mg/dL lower LDL, 2.5 kg lower weight, and 0.37 percentage points lower HbA1c.

Full veganism is one way to do this, not the only evidence-supported route. A plant-rich Mediterranean pattern has randomized evidence over 18 months, and well-executed medical diets remain credible comparators. Anyone adopting a fully vegan diet should plan for vitamin B12, vitamin D, omega-3 fats, and zinc rather than treating supplementation as an afterthought. Monitor lipids, glucose, weight, and any medication needs with a clinician—especially if you use glucose-lowering or blood-pressure drugs. The diet can move biomarkers; the hard-outcome trial has yet to arrive.