Two Studies, Two Conclusions?

This article frames the debate over ketogenic and carnivore-style diets by contrasting two very different studies: a large observational analysis showing higher cardiovascular risk and a small imaging trial suggesting no short-term plaque progression despite high ApoB. The discussion matters because patients, clinicians, and researchers need to balance population-level associations with targeted mechanistic data. Only by comparing evidence of different strengths and limitations can we move toward clearer guidance on the long-term safety of very low-carbohydrate diets.

The two studies approach the same dietary question from very different angles. The ACC analysis drew from the UK Biobank, identifying people who self-reported eating “keto-like” diets (low carbohydrate, high fat) and tracking them for about 12 years. Researchers found that these individuals had higher LDL cholesterol and ApoB levels, and nearly twice the rate of cardiovascular events compared to matched controls. Its strength lies in the large sample size and long follow-up, but because diet was self-reported and the design was observational, it cannot establish causation. Confounding variables such as lifestyle differences, food quality, or reporting bias could have influenced the outcomes.
The Keto-CTA trial, led by Norwitz and colleagues, examined 100 individuals who met a strict lean mass hyper-responder profile: very high LDL, high HDL, low triglycerides, all diet-induced on a ketogenic diet. Over one year, participants underwent coronary CT angiography, a direct measure of plaque burden. The study found no link between ApoB/LDL levels and plaque progression during that short follow-up; baseline plaque burden was the main predictor of change. This design used objective imaging and carefully defined metabolic criteria, but its limitations include small size, short duration, and a highly selective, healthy cohort.

Dr. Layne Norton, a nutrition scientist and coach, has publicly cautioned against overinterpreting the Keto-CTA results. He points out that a one-year study, especially in a highly screened population, cannot overturn decades of research linking elevated ApoB to higher cardiovascular risk. Norton emphasizes that plaque progression is a slow process, often taking years or decades to manifest, and that the absence of detectable change over 12 months does not prove safety in the long run. He argues that while mechanistic trials are valuable, they must be weighed against the large body of prospective and genetic evidence connecting LDL and ApoB to heart disease.

I find even less value in the Keto-CTA trial. Several of the authors themselves have known leanings toward ketogenic or carnivore-style diets, which raises legitimate concerns that the study may have been conducted in a way that fits, rather than challenges, their pre-conceptions. For example:

Nicholas G. Norwitz has a history of publishing pro–low-carb work and co-authored a Mediterranean low-carbohydrate cookbook, reflecting long-standing support for carb-restricted diets.

David Feldman is a co-founder of the Citizen Science Foundation (CSF), the very organization that funded the trial. CSF’s mission is centered on the “Lean Mass Hyper-Responder” (LMHR) phenomenon—a profile that emerges almost exclusively in ketogenic communities—and Feldman is an active public figure in the keto space.

Thomas R. Wood and Adrian Soto-Mota have both previously published work sympathetic to low-carb dietary strategies.

The fact that the study was funded through CSF, an organization deeply embedded in the keto community, combined with authors who are publicly identified as keto or carnivore proponents, makes it reasonable to question the neutrality of the trial. It could appear that the authors entered and published this study not to challenge their assumptions, but to confirm a pre-conceived narrative: that elevated ApoB in keto-adapted individuals is harmless. Methodologically, the trial relied on a very small, highly selective sample of LMHRs with unusually favorable metabolic markers, making the results difficult to generalize to broader populations. The absence of a well-defined control group and the one-year follow-up—which is far too short to capture the decades-long development of atherosclerosis—further limit its value. At best, it provides a mechanistic snapshot, but it adds little clarity to the real-world question of whether very low-carbohydrate, meat-heavy diets improve or worsen long-term cardiovascular outcomes.

Comparison: ACC “Keto-Like” Observational Study vs. Norwitz Keto-CTA Trial

Study	ACC / UK Biobank Observational (“Keto-Like” Diet)	Keto-CTA Trial (Norwitz et al.)
Design	Observational, dietary intake self-reported; long-term follow-up using UK Biobank data.	Prospective interventional follow-up; baseline and 1-year CT angiography with labs.
Sample Size	~305 people with “keto-like” diet matched to ~1,200 controls (from ~70,000 total).	100 participants, all lean mass hyper-responders (LMHR).
Duration	Median ~12 years follow-up for CVD events.	1 year imaging follow-up (some had been keto-adapted ~4.7 years before enrollment).
Inclusion	Self-reported “low-carb, high-fat” (≤25% carb, >45% fat).	LDL-C ≥190 mg/dL, HDL ≥60 mg/dL, TG ≤80 mg/dL, ketogenic diet adherent, metabolically healthy.
Exclusion	Not specifically detailed beyond dietary reporting exclusions.	Prior CVD, diabetes, uncontrolled hypertension, lipid-lowering therapy, FH, inability for CT imaging.
Endpoints	Major cardiovascular events (MI, stroke, revascularization, angina, blocked arteries).	Plaque volume, noncalcified plaque, total plaque score progression by CT angiography.
Main Findings	“Keto-like” group had higher LDL/ApoB and nearly double the risk of CVD events.	No association between ApoB/LDL and plaque progression over 1 year in LMHRs; baseline plaque predicted progression.
Strengths	Large sample, long follow-up, real-world outcomes.	Direct imaging of arteries; objective biomarkers; tightly defined metabolic profile.
Limitations	Self-reported diet; broad definition of “keto-like”; confounding; not randomized.	Small sample; only 1 year follow-up; selected healthy subgroup; unknown long-term outcomes.

Taken together, these studies show the tension between breadth and precision. The ACC analysis highlights broad, long-term associations that raise legitimate concerns, while the Keto-CTA trial provides narrowly focused, short-term mechanistic data in a highly selective group. Layne Norton is right to caution that one year of imaging cannot outweigh decades of evidence connecting ApoB to cardiovascular risk. And from my perspective, the Norwitz study—authored largely by low-carb proponents, funded by a foundation tied to the keto community, limited in scope, short in duration, and lacking a proper control group—reads more like confirmation of a belief system than a neutral test of safety. The weight of evidence still leans toward caution, and meaningful guidance will only come from large, long-term, independently funded trials that examine diverse populations and capture the complexity of diet, metabolism, and cardiovascular health.