Keto vs. Whole Food Plant-Based: A Deep Dive Into Cardiovascular Evidence

Two dietary philosophies dominate the cardiovascular prevention conversation right now: ketogenic (very low carbohydrate, high fat) and whole food, low-fat plant-based (WFPB). Both have passionate advocates, both have real evidence behind them, and both produce genuinely different biological effects. This post tries to give an honest, neutral look at where the evidence actually stands, going well beyond the standard lipid panel.

Before anything else: both diets exist on a wide spectrum of quality. A keto diet built around grass-fed beef, fatty fish, eggs, avocado, nuts, and non-starchy vegetables is a fundamentally different biological intervention than one built around processed meats, cheese, and keto snack bars. A WFPB diet centered on legumes, intact whole grains, vegetables, fruits, nuts, and seeds is categorically different from a low-fat vegan diet heavy in refined carbohydrates, white rice, and processed plant foods. This discussion applies only to the healthiest, most nutrient-complete versions of each approach.

For plant-based eaters, the nutrients needing consistent attention are well-established: vitamin B12 (essentially absent from plant foods and non-negotiable to supplement), vitamin D, the long-chain omega-3s EPA and DHA (conversion from ALA is inefficient in many people, making algae-based supplementation worthwhile), zinc, iodine, and for some people calcium and iron. These gaps are well-recognized and generally well-managed in informed plant-based communities.

The nutrient concern list for keto is less widely talked about but equally real, and several items carry direct cardiovascular relevance. Folate is a significant one, and legumes and fortified grains are primary sources and both are eliminated on keto, and folate deficiency drives homocysteine elevation, which is directly toxic to the endothelium. Potassium is another real concern: fruits, legumes, and whole grains are major dietary sources, all restricted, and keto's well-documented urinary electrolyte losses make this worse. Vitamin C is heavily concentrated in fruits and starchy vegetables, and chronic subclinical deficiency is plausible on poorly constructed keto, which is relevant because vitamin C is an eNOS cofactor and important for vascular collagen integrity. Fiber, while not a classical nutrient, is functionally critical, and most keto practitioners fall well short of adequate amounts, with downstream effects on microbiome diversity, butyrate production, and systemic inflammation. Additional concerns include thiamine (B1, lost with whole grain elimination), manganese (concentrated in grains and legumes), vitamin K1 (if non-starchy vegetable intake is low), and calcium (if dairy is limited or avoided). This discussion assumes both groups are thoughtfully addressing these gaps through food selection and supplementation, so deficiency is not a confounding variable in the cardiovascular comparison.

The Standard Lipid Picture

The lipid phenotypes these two diets produce are genuinely different and both internally coherent. Well-formulated keto tends to produce lower triglycerides, higher HDL, and often lower small dense LDL, driven largely by reduced insulin signaling, suppressed hepatic de novo lipogenesis, and increased fat oxidation. Low-fat WFPB tends to produce lower LDL and ApoB, but frequently higher triglycerides and lower HDL, a consequence of higher carbohydrate intake stimulating hepatic VLDL secretion even from healthy whole food sources.

ApoB, the particle count measure that most directly reflects atherogenic burden, tends to favor the plant-based approach, particularly on very low fat versions of the diet. On keto, ApoB is highly variable: many people see modest reductions, but a meaningful subset, sometimes called lean mass hyper-responders, experience dramatic LDL and ApoB elevations despite optimal TG and HDL. This remains one of the more clinically concerning and unresolved findings in the keto literature.

The triglyceride elevation seen on low-fat plant-based diets is real and worth monitoring, but context matters. Triglycerides driven by high intake of whole food carbohydrates in a low-fat, low-ApoB metabolic environment are a different pathophysiological signal than triglycerides arising from insulin resistance and metabolic syndrome. Whether they carry equivalent remnant cholesterol risk is genuinely uncertain.

Beyond Lipids: Pleiotropic Effects

This is where the comparison gets considerably more complex, and where the standard lipid panel tells only a fraction of the story.

A well-designed ketogenic diet can and should include meaningful quantities of non-starchy vegetables, leafy greens, broccoli, cauliflower, asparagus, zucchini, peppers, and others, along with nuts, seeds, avocado, olive oil, and fatty fish. These foods are not trivial: they contribute polyphenols, vitamin K1, some dietary nitrates, fiber (modest relative to plant-based but not zero), marine omega-3s, and monounsaturated fats with established endothelial benefits. A keto diet built thoughtfully around these foods is genuinely different from one built around processed meats and cheese, and its pleiotropic profile is meaningfully better than critics sometimes acknowledge.

That said, the breadth, volume, and variety of phytonutrient-rich foods available on a whole food plant-based diet is categorically wider. Legumes, intact whole grains, a full spectrum of fruits, root vegetables, and starchy vegetables, all absent or severely restricted on keto, are among the most potent sources of specific polyphenol classes, prebiotic fibers, resistant starch, isoflavones, lignans, and other bioactive compounds with direct vascular relevance. Some of these foods and their constituent compounds simply cannot be replicated within keto constraints regardless of how carefully the diet is constructed.

This distinction becomes especially significant when the goal is active vascular repair rather than just slowing progression. The endothelial lining is a dynamic, regenerative tissue. Endothelial progenitor cells, circulating cells responsible for repairing damaged vascular endothelium, are mobilized and supported by specific dietary signals, including nitric oxide availability, VEGF signaling, and reduced oxidative stress. Dietary nitrates from beets, leafy greens, and other plant sources drive eNOS activity and NO bioavailability in ways that support both acute vasodilation and longer-term endothelial regeneration. The polyphenol load achievable on a full plant-based diet, particularly from berries, pomegranate, cocoa, green tea, and legumes, activates Nrf2-mediated antioxidant gene expression, suppresses NF-κB-driven inflammatory gene transcription, and supports sirtuin pathways associated with vascular protection and cellular repair. These are not marginal effects, they represent a coordinated shift in vascular gene expression toward protection and away from inflammation and dysfunction.

A well-designed keto diet can access some of these mechanisms, such as NO support from leafy greens, Nrf2 activation from cruciferous vegetables, anti-inflammatory signaling from omega-3s. But the magnitude, the variety of pathways engaged simultaneously, and the inclusion of uniquely potent whole food sources that keto simply cannot accommodate means the overall pleiotropic vascular effect is likely substantially greater on a whole food plant-based diet, and this gap is probably most consequential in people with already-damaged endothelium who need the most robust biological support for repair and regeneration.

High fiber intake on plant-based diets robustly shifts the gut microbiome toward short-chain fatty acid producers with downstream anti-inflammatory endothelial effects. Critically, plant-based diets dramatically suppress TMAO production, not only by reducing choline and carnitine substrates from animal foods, but by reshaping the microbiome away from TMA-producing organisms. TMAO is increasingly credible as an independent endothelial toxin and pro-thrombotic signal. Notably, a plant-based person who occasionally eats eggs produces far less TMAO than an omnivore eating the same eggs, because the microbiome itself is different.

Plant-based diets also deliver gram-level phytosterols daily, which competitively inhibit cholesterol absorption, and eliminate heme iron, a pro-oxidant that catalyzes lipid peroxidation and promotes oxidized LDL formation through Fenton chemistry at a level that non-heme iron from plants does not.

On the keto side, the elimination of refined carbohydrates and added sugars is a genuinely powerful pleiotropic benefit, particularly for metabolically unhealthy individuals. Reduced fasting insulin, lower glucose variability, and decreased glycation end-product formation all translate to meaningful reductions in endothelial damage, through pathways that a plant-based diet addresses differently and potentially less efficiently in insulin-resistant phenotypes. Emerging data also suggests that beta-hydroxybutyrate may have direct anti-inflammatory effects through NLRP3 inflammasome inhibition and epigenetic mechanisms, though whether dietary ketosis produces sufficient circulating levels for meaningful cardiovascular pleiotropic effects in practice remains an open question.

Keto diets high in red meat, however, carry real pleiotropic costs. Saturated fatty acids, particularly palmitic acid, activate inflammatory signaling on endothelial cells and macrophages, acutely suppress nitric oxide production, and produce measurable reductions in flow-mediated dilation in controlled feeding studies. Heme iron loading from red meat drives oxidative stress in arterial tissue. And in an omnivore microbiome, the high choline and carnitine content of animal-rich diets sustains TMAO production at levels plant-based diets do not, a cost that favorable TG and HDL numbers may not offset.

Beyond these better-known mechanisms, regular consumption of red meat and other animal foods introduces additional biologically active compounds worth considering in a cardiovascular context. Neu5Gc (N-glycolylneuraminic acid) is a non-human sialic acid found abundantly in mammalian meat and dairy that humans cannot synthesize but do incorporate into tissues after dietary exposure. Because humans produce antibodies against Neu5Gc, this incorporation triggers a chronic low-grade inflammatory response termed xenosialitis, a mechanism with direct relevance to endothelial inflammation and atherosclerosis that operates entirely independently of lipid levels. Animal research has demonstrated accelerated atherosclerosis from this mechanism, and while human data remains limited, the biological plausibility is significant and the exposure is chronic and cumulative in high red meat consumers.

Additionally, animal foods, particularly those processed through conventional slaughter and handling, carry a meaningful burden of bacterial endotoxin (lipopolysaccharide, or LPS) from gram-negative intestinal bacteria. Dietary LPS is not fully neutralized by cooking and is absorbed to varying degrees across the gut epithelium, where it activates TLR4 receptors on endothelial cells and circulating monocytes, driving systemic inflammatory signaling. Postprandial endotoxemia following animal-food-rich meals has been documented in controlled studies and is thought to contribute to the chronic low-grade inflammatory state associated with Western dietary patterns. A high-fat meal context, as is typical on keto, may facilitate LPS absorption via chylomicron incorporation, potentially amplifying this effect. Plant foods do not carry this endotoxin burden, and high-fiber plant-based diets appear to support gut barrier integrity in ways that further reduce systemic LPS translocation.

Taken together, Neu5Gc-driven xenosialitis and dietary endotoxemia represent two additional and underappreciated pathways by which regular red meat and animal food consumption may actively impede vascular healing, sustain endothelial inflammation, and contribute to plaque development and instability, regardless of how the standard lipid panel looks.

Plaque: The Outcome That Actually Matters

The evidence on arterial plaque is asymmetric, and that asymmetry deserves honest acknowledgment.

The KETO-CTA study, the only prospective coronary CT imaging data in a ketogenic population, examined lean mass hyper-responders and initially reported plaque progression as a predominant finding despite favorable triglyceride and HDL phenotypes. The keto community has contested the methodology vigorously, and an ongoing reevaluation of the data is now revealing a more nuanced picture: a subset of participants appears to show plaque regression or stabilization, which investigators and keto advocates have highlighted as an encouraging signal. That is worth acknowledging. However, even in the revised analysis, plaque progression still appears to be more common than regression across the cohort, and the full distribution of individual responses continues to be worked through publicly in real time.

More fundamentally, the study's design limits how much any finding, in either direction, can reasonably be interpreted. It is a single-arm, non-randomized observational study with no control group and only one year of follow-up, far too short to assess cardiovascular events, mortality, or reliable longer-term plaque trajectory. Without a concurrent control group eating a different diet under similar conditions, it is not possible to separate diet-specific effects from natural disease progression, regression to the mean, or other confounders. A subset showing regression in an uncontrolled one-year imaging study is an interesting hypothesis-generating observation, not evidence of benefit. The study cannot currently be placed anywhere near the same evidentiary tier as controlled dietary intervention trials with hard endpoints, and representing it as proof that keto is safe or beneficial for coronary plaque would significantly overreach what the data can actually support at this stage.

On the plant-based side, Ornish's Lifestyle Heart Trial demonstrated angiographic regression of coronary stenosis alongside significant reductions in cardiac events, and remains one of the most cited pieces of evidence in this space. It deserves that status, but its limitations also deserve honest acknowledgment. The imaging technique used, quantitative coronary angiography, measures luminal stenosis rather than total plaque volume or composition, and by today's standards it is a relatively crude tool compared to CCTA with plaque characterization, IVUS, or OCT. The intervention was also intensive and multifactorial, combining a very low fat plant-based diet with stress management, exercise, group support, and smoking cessation, making it genuinely difficult to attribute the observed regression specifically to diet alone versus the full lifestyle package. The study was also small, and dropout and selection effects in such demanding trials are real concerns. None of this invalidates the findings, but it does mean the Ornish data is best understood as strong evidence for an intensive lifestyle intervention that included a plant-based diet, rather than clean proof of what the diet does in isolation.

Esselstyn's work, showing remarkable arrest of progression and in some cases apparent regression in patients with severe established CAD, is compelling and clinically meaningful, particularly given the severity of disease in his cohort and the durability of outcomes in adherent patients. The critiques here are also legitimate: it is an uncontrolled case series without randomization or a comparison group, the imaging assessments were not blinded, the populations were highly selected and motivated, and the intervention again combined dietary change with close clinical follow-up and support that could independently influence outcomes. As with Ornish, adherence was the dominant predictor of outcome, which is both encouraging and a reminder that these results reflect what happens in people who commit fully, not average real-world adherence.

Inflammation markers, such as CRP, IL-6, oxidized LDL, fall substantially on well-designed plant-based interventions, and these reductions likely influence plaque morphology in ways that imaging of size alone does not capture: a lower-inflammation plaque environment is thought to produce thicker fibrous caps and smaller lipid cores, reducing vulnerability to rupture and MACE risk independent of stenosis degree.

It is also worth noting that in Ornish's original trial, triglycerides rose modestly and HDL fell on the very low fat plant-based diet, and yet regression occurred. This suggests that the cardiovascular benefit of the diet operates through mechanisms the TG/HDL phenotype does not fully capture.

Individual Variation: A Factor That Cannot Be Ignored

Both dietary approaches fail a subset of people in ways that matter clinically, and intellectual honesty requires acknowledging this directly.

Some individuals on well-formulated plant-based diets struggle to meet protein targets without substantially increasing carbohydrate load or relying on more refined foods, experience persistent fatigue, or find that their specific cardiovascular markers, particularly ApoB or particle number, do not respond as favorably as expected. For these individuals, modifications make sense: shifting toward somewhat higher plant fat intake from nuts, seeds, and avocado; emphasizing protein-dense plant foods like tempeh, edamame, or lentils more deliberately; or modestly reducing total carbohydrate while keeping the diet plant-centered. None of these adjustments require abandoning the core framework.

Some individuals on ketogenic diets experience energy limitations, poor exercise recovery, mood effects, or, importantly, unfavorable ApoB and LDL trajectories that do not self-resolve over time. For these individuals, adding strategic carbohydrate from whole food sources, or moderating saturated fat while preserving low overall carbohydrate intake, may be necessary to achieve a metabolic and cardiovascular risk profile that actually serves them. The diet should serve the person, not the other way around.

Both communities have a tendency toward ideological rigidity that does not serve patients or the science. Individual genomics, gut microbiome composition, baseline metabolic health, thyroid function, exercise demands, and other factors create genuine biological heterogeneity in dietary response. Monitoring actual biomarkers, ApoB, inflammatory markers, glucose, and where accessible, imaging, rather than assuming dietary adherence alone guarantees a favorable outcome, is essential on either approach.

The Question of "How Plant-Based Is Plant-Based Enough?"

For those interested in the plant-based direction, an important nuance: the evidence does not require perfect adherence to produce meaningful cardiovascular benefit. And some of the most instructive data here actually comes from studies that were neither vegan nor particularly ideologically motivated, just well-designed dietary intervention trials that happen to shed light on this question.

The PREDIMED trial, a large, well-designed randomized trial, demonstrated significant reductions in major cardiovascular events in participants eating a Mediterranean diet supplemented with olive oil or nuts. This was not a vegan diet; it included fish, poultry, eggs, and some dairy. Yet it produced robust cardiovascular protection, and subsequent imaging substudies have shown favorable effects on carotid atherosclerosis. The Lyon Diet Heart Study similarly showed dramatic reductions in recurrent cardiac events on a modified Mediterranean pattern, again, not plant-exclusive, but heavily plant-forward and low in saturated fat. More recently, several CIMT and plaque imaging studies using predominantly plant-based but not strictly vegan dietary patterns have shown regression or stabilization of subclinical atherosclerosis, adding to the picture that the cardiovascular benefit of moving toward whole plant foods operates on a continuum rather than requiring a binary threshold.

Taken together, the large cohort data from the Adventist Health Studies and others, combined with these intervention trials, suggests a consistent dose-response relationship: the more closely one approximates a whole food plant-based pattern, displacing animal products and processed foods in favor of vegetables, legumes, fruits, whole grains, nuts, and seeds, the more favorable the cardiovascular biomarker and event data becomes. But meaningful benefit appears well before 100% adherence, and the gap between a 90% and 100% plant-based diet is almost certainly smaller than the gap between either of those and a standard Western diet. Pursuing the most plant-rich version of a diet that is genuinely sustainable for a given individual is likely more valuable than perfect theoretical adherence that cannot be maintained in real life.

Summary

Both diets, at their best, represent significant improvements over the standard Western dietary pattern. Both produce real cardiovascular benefits through distinct and partially overlapping mechanisms. The keto approach has particular strength in metabolically unhealthy, insulin-resistant individuals where glucose and insulin normalization may be the dominant cardiovascular risk driver. The plant-based approach has a broader pleiotropic cardiovascular profile, covering endothelial nitric oxide biology, TMAO suppression, plaque inflammation and morphology, polyphenol-mediated endothelial protection, and phytosterol effects, and is the only dietary pattern with prospective evidence of coronary plaque regression and hard event reduction in established ASCVD.

The honest summary is that the evidence base is stronger and broader for plant-based diets in cardiovascular outcomes, but the keto approach is not without at least some real benefit at least in terms of cardiometabolic biomarkers and certain disease states outside of ASCVD, though the evidence in terms of bottom line plaque regression and stabilization is far from adequate to state definitively that it will help there, while the evidence in terms of event reduction is absent at this point. Nonetheless, neither approach is universally optimal for every single human being, and both require thoughtful individualization. No matter what you choose, and what I, or any study suggest may or may not happen from doing so, the bottom line is YOUR BODY and how it responds. Therefore my simplest advice of all is, no matter what you choose, please be sure to test and don't guess on the impact your choices are having for your body. If you would like to explore things more for yourself, and better evaluate your personal situation, and/or talk through your dietary and lifestyle options, supplements, medications, or all of the above, please don't delay. There is no time like the present for best supporting your arteries. You can schedule an in person (Portland, Oregon) or virtual appointment with me at THIS LINK.