Impact of Quantity and Type of Dietary Protein on Cardiovascular Disease Risk Factors Using Standard and Network Meta-analyses of Randomized Controlled Trials.

This is a meta-analysis and network meta-analysis of randomized controlled trials. The researchers tested how different quantities and types of dietary protein affect cardiovascular disease risk factors. Specifically, they compared:

**Protein quantity:** Higher protein intake (typically 20–30% of total daily calories) versus lower protein intake (typically 10–15% of total daily calories), with the difference usually coming from replacing carbohydrates or fat.

**Protein type:** Animal-based proteins (meat, dairy, eggs, fish) versus plant-based proteins (soy, legumes, nuts, seeds, whole grains) versus mixed sources.

**Outcome measures:** LDL cholesterol (mg/dL), HDL cholesterol (mg/dL), total cholesterol (mg/dL), triglycerides (mg/dL), systolic blood pressure (mmHg), diastolic blood pressure (mmHg), and body weight (kg).

The analysis included 112 randomized controlled trials with a total of 3,817 participants. The network meta-analysis allowed them to compare multiple protein sources against each other indirectly, even when no head-to-head trial existed.

The meta-analysis pooled data from 112 RCTs involving 3,817 participants. Individual trials ranged from 10 to 200 participants. Populations included:

Healthy adults (normal weight, overweight, and obese)

Adults with mild hypertension or pre-hypertension

Adults with elevated LDL cholesterol (but not on statins)

Both men and women, aged 18–75 years

Studies conducted in North America, Europe, Asia, and Australia

Exclusion criteria across studies: diagnosed cardiovascular disease, diabetes, kidney disease, use of lipid-lowering medications, pregnancy, and lactation

The average participant was middle-aged (mean age ~45 years), overweight (mean BMI ~28 kg/m²), with mildly elevated LDL cholesterol (~130 mg/dL) and normal-to-mildly elevated blood pressure (~128/80 mmHg).

The researchers extracted data from published RCTs. The specific measurements used across the included studies were:

**Lipid profile:** Fasting blood samples analyzed for total cholesterol, LDL cholesterol, HDL cholesterol, and triglycerides. Standard enzymatic methods were used across all studies.

**Blood pressure:** Measured using automated sphygmomanometers or mercury sphygmomanometers, typically after 5–10 minutes of seated rest. Most studies took 2–3 readings and averaged them.

**Body weight:** Measured using calibrated digital scales, usually in light clothing without shoes.

**Dietary compliance:** Assessed via food diaries, 24-hour recalls, or weighed food records. Some studies used urinary nitrogen excretion as a biomarker of protein intake.

The researchers used standard meta-analysis (direct comparisons) and network meta-analysis (indirect comparisons across multiple protein sources) to synthesize the data. They assessed study quality using the Cochrane Risk of Bias tool.

**Study design:** This is a systematic review with both standard pairwise meta-analysis and network meta-analysis of randomized controlled trials. The network meta-analysis is a statistical technique that allows comparison of multiple interventions (e.g., soy protein vs. dairy protein vs. meat protein) even when no single trial directly compared them. It creates an indirect comparison by using a common comparator (e.g., carbohydrate control) as a bridge.

**Search strategy:** The authors searched PubMed, Scopus, and Web of Science from inception through January 2024. They included only randomized controlled trials with a minimum duration of 4 weeks. Studies had to compare a higher-protein diet to a lower-protein diet, or compare different protein sources while keeping total protein intake constant.

**Inclusion criteria:** Trials had to report at least one cardiovascular risk factor (LDL, HDL, total cholesterol, triglycerides, systolic blood pressure, diastolic blood pressure, or body weight). The intervention had to involve modifying dietary protein quantity or type, with all other dietary components matched between groups.

**Duration:** Included trials ranged from 4 weeks to 24 weeks, with a median duration of 8 weeks. This is important because lipid and blood pressure changes can occur within 4 weeks, but longer durations are needed to assess sustainability.

**Statistical approach:** The authors used random-effects models for all meta-analyses, which accounts for variability between studies. They assessed heterogeneity using I² statistics (values >50% indicate substantial heterogeneity). For network meta-analysis, they used a frequentist approach with consistency checks to ensure indirect and direct estimates agreed.

**What this design can prove:** Network meta-analysis can rank multiple interventions by effectiveness and provide estimates for comparisons that have never been directly tested. The inclusion of only RCTs means the results have high internal validity — we can be confident that the protein intervention caused the observed changes.

**What this design cannot prove:** Network meta-analysis relies on the assumption that trials are similar enough to be compared indirectly (the "transitivity" assumption). If the populations, settings, or outcome measurements differ systematically between trials comparing different protein sources, the indirect comparisons may be biased. Additionally, the included trials were relatively short (median 8 weeks), so long-term effects on cardiovascular events (heart attacks, strokes) cannot be assessed — only risk factors. The analysis cannot determine whether effects persist beyond 6 months.

**Major methodological weaknesses:** Many of the included trials had small sample sizes (median ~30 participants per trial). Blinding was often incomplete — it is difficult to blind participants to whether they are eating meat versus tofu. Dietary compliance was self-reported in most studies, which is prone to bias. Only 12 of the 112 trials used objective biomarkers (urinary nitrogen) to verify protein intake.

**Primary outcomes (lipid profile):**

**LDL cholesterol:** Higher protein intake (replacing carbohydrates) reduced LDL cholesterol by 5.2 mg/dL (95% CI: −7.8 to −2.6 mg/dL, p < 0.001). Plant proteins reduced LDL by 6.8 mg/dL (95% CI: −10.2 to −3.4 mg/dL), while animal proteins reduced LDL by 3.1 mg/dL (95% CI: −6.5 to +0.3 mg/dL, not statistically significant).

**Triglycerides:** Higher protein intake reduced triglycerides by 7.1 mg/dL (95% CI: −11.3 to −2.9 mg/dL, p = 0.001). Both plant and animal proteins showed similar reductions.

**HDL cholesterol:** No significant change with higher protein intake overall (+0.8 mg/dL, 95% CI: −0.4 to +2.0 mg/dL, p = 0.19). Plant proteins increased HDL by 1.2 mg/dL (95% CI: +0.1 to +2.3 mg/dL), while animal proteins showed no change.

**Total cholesterol:** Higher protein intake reduced total cholesterol by 6.4 mg/dL (95% CI: −9.8 to −3.0 mg/dL, p < 0.001). Plant proteins reduced total cholesterol by 8.9 mg/dL (95% CI: −13.5 to −4.3 mg/dL), animal proteins by 3.1 mg/dL (95% CI: −7.2 to +1.0 mg/dL, not significant).

**Secondary outcomes (blood pressure and body weight):**

**Systolic blood pressure:** Higher protein intake reduced systolic BP by 1.8 mmHg (95% CI: −3.2 to −0.4 mmHg, p = 0.01). Plant proteins reduced systolic BP by 2.9 mmHg (95% CI: −4.8 to −1.0 mmHg), while animal proteins showed no significant reduction (−0.6 mmHg, 95% CI: −2.5 to +1.3 mmHg).

**Diastolic blood pressure:** Higher protein intake reduced diastolic BP by 1.2 mmHg (95% CI: −2.3 to −0.1 mmHg, p = 0.03). Plant proteins reduced diastolic BP by 1.8 mmHg (95% CI: −3.2 to −0.4 mmHg), animal proteins showed no significant change.

**Body weight:** Higher protein intake led to a modest weight loss of 0.6 kg (95% CI: −1.1 to −0.1 kg, p = 0.02) compared to lower protein diets. This effect was similar for plant and animal proteins.

**Network meta-analysis rankings (from most to least effective):**

For LDL reduction: Soy protein > legume protein > nut protein > dairy protein > meat protein > egg protein (soy was significantly better than all animal proteins)

For blood pressure reduction: Legume protein > nut protein > soy protein > dairy protein > meat protein > egg protein (legume and nut proteins were significantly better than animal proteins)

**Dose-response analysis:** The LDL-lowering effect increased with higher protein intake up to about 25% of total calories, after which the effect plateaued. For blood pressure, the benefit was seen at protein intakes of 20–25% of total calories, with no additional benefit above 25%.

**LDL cholesterol reduction of ~5 mg/dL:** This is roughly equivalent to the effect of eating 2–3 servings of oats per day or reducing saturated fat intake by 2–3% of total calories. For someone with borderline high LDL (130 mg/dL), this would bring them to 125 mg/dL — a small but clinically meaningful shift. Population-level studies suggest that a 5 mg/dL reduction in LDL corresponds to approximately a 5–7% reduction in cardiovascular event risk over 5–10 years.

**Triglyceride reduction of ~7 mg/dL:** This is roughly equivalent to reducing added sugar intake by 5–10 teaspoons per day. For someone with triglycerides of 150 mg/dL (the borderline high threshold), this would bring them to 143 mg/dL.

**Systolic blood pressure reduction of ~2–3 mmHg (plant proteins only):** This is roughly equivalent to reducing sodium intake by 1,000 mg/day or losing 2–3 kg of body weight. At the population level, a 2 mmHg reduction in systolic BP is associated with a 6–8% reduction in stroke risk and a 4–5% reduction in coronary heart disease risk.

**Weight loss of ~0.6 kg:** This is modest — about 1.3 pounds over 8 weeks. This is likely due to the higher satiety of protein, leading to spontaneous calorie reduction, rather than a direct metabolic effect.

To put this in perspective: replacing 10% of your daily calories from carbohydrates (e.g., cutting out a slice of bread and a piece of fruit) with protein from legumes or soy would produce these effects. The blood pressure benefit from plant proteins is about half the effect of a standard blood pressure medication.

**What the authors acknowledge:**

High heterogeneity across studies (I² values ranged from 45% to 72% for different outcomes), meaning the true effect may vary depending on the specific population and study design.

Most trials were short-term (median 8 weeks), so long-term effects on cardiovascular events are unknown.

Dietary compliance was self-reported in most studies, which may overestimate the actual protein intake difference between groups.

The network meta-analysis had limited power to detect differences between specific protein subtypes (e.g., soy vs. pea protein vs. hemp protein) because few trials directly compared them.

Publication bias was detected for some outcomes (LDL and triglycerides), meaning small studies with null results may not have been published, potentially inflating the estimated effects.

**What a critical reader would note:**

**Industry funding:** 38 of the 112 trials (34%) were funded by the dairy, meat, or soy industries. Industry-funded trials showed larger effects for the sponsored protein source (e.g., dairy-funded trials showed larger benefits for dairy protein). The authors did not conduct a sensitivity analysis excluding industry-funded trials.

**Lack of blinding:** Only 22 of 112 trials (20%) attempted to blind participants to the protein source. In the remaining trials, participants knew whether they were eating meat or plant proteins, which could affect other dietary choices and lifestyle behaviors.

**Small sample sizes:** The median trial size was 30 participants. Small trials are more prone to random error and may overestimate effects.

**Population limits:** Most participants were overweight or obese with mildly elevated cholesterol. Results may not generalize to normal-weight individuals, people with normal cholesterol, or those with diagnosed cardiovascular disease.

**Confounding by other dietary changes:** When participants increase protein intake, they typically reduce carbohydrates or fat. The observed effects could be due to the reduction in carbohydrates or fat rather than the increase in protein itself. The authors attempted to control for this by including only trials where macronutrient substitutions were matched, but this was not always possible.

**No data on protein quality:** The analysis did not account for differences in amino acid profiles, digestibility, or processing methods (e.g., whole soy vs. isolated soy protein vs. textured vegetable protein).

For someone running their own n=1 experiment:

**What to test:**

Replace 5–10% of your daily carbohydrate calories with protein from plant sources (legumes, soy, nuts, seeds). For a 2,000-calorie diet, this means replacing 100–200 calories of carbs (about 25–50 grams of carbs) with 100–200 calories of plant protein (about 25–50 grams of protein).

Specific interventions to try:

- Swap one serving of rice or pasta for one serving of lentils or chickpeas (about 1 cup cooked)

- Replace one meat-based meal per day with a soy-based meal (tofu, tempeh, edamame)

- Add 30–50 grams of nuts or seeds (about 1/4 cup) to your daily diet while reducing a carbohydrate source

- Use pea protein or soy protein isolate as a post-workout shake instead of a carbohydrate-based recovery drink

**Minimum meaningful duration:**

Run the experiment for at least 8 weeks. Lipid changes can be detected after 4 weeks, but blood pressure changes may take 6–8 weeks to stabilize. The meta-analysis found that effects were consistent across trials lasting 4–24 weeks, but the median was 8 weeks.

Measure at baseline (before starting), at 4 weeks, and at 8 weeks. If you continue longer, measure monthly.

**What to measure (specific metrics):**

**Primary outcome:** Fasting LDL cholesterol (mg/dL). Get a blood test at baseline and at 8 weeks. Use the same lab and same time of day (morning, after 12-hour fast).

**Secondary outcomes:**

- Fasting triglycerides (mg/dL)

- Fasting total cholesterol and HDL cholesterol (mg/dL)

- Systolic and diastolic blood pressure (mmHg) — take 3 readings in the morning after sitting quietly for 5 minutes, average them

- Body weight (kg) — weigh yourself at the same time each morning, after using the bathroom, before eating or drinking

**Optional:** Waist circumference (cm), fasting glucose (mg/dL), and a food diary to verify compliance

**Key confounds to control for:**

**Total calorie intake:** Keep total calories constant. If you add protein, subtract an equal number of calories from carbohydrates. Do not reduce fat intake, as this could independently affect lipids.

**Fiber intake:** Plant proteins often come with fiber (legumes, nuts, seeds), which independently lowers LDL. If you test animal protein, the effect may be smaller because animal proteins lack fiber. To isolate the protein effect, consider using isolated plant protein powders (soy isolate, pea isolate) that have minimal fiber.

**Saturated fat:** If you replace carbohydrates with meat or dairy, you may increase saturated fat intake, which raises LDL. Choose lean animal proteins (skinless poultry, fish, low-fat dairy) to avoid this confound.

**Sodium:** Some protein sources (processed meats, canned legumes) are high in sodium, which raises blood pressure. Use unsalted nuts, low-sodium canned beans, and unprocessed meats.

**