Is high LDL always dangerous — the Lean Mass Hyper-Responder challenge

The apoB-causes-heart-disease model is one of the most thoroughly tested ideas in cardiovascular medicine, supported by Mendelian randomization studies, statin trials, and PCSK9 inhibitor data [1]. A small but vocal research community has spent several years asking whether that model applies universally, or whether a specific phenotype applies different rules: lean people whose LDL soars on ketogenic diets. The first real prospective data on that question arrived in 2025.

The Lean Mass Hyper-Responder phenotype

Lean Mass Hyper-Responders (LMHRs) are individuals who, upon adopting a low-carbohydrate diet, develop strikingly improved LDL alongside low triglycerides and high HDL. Dave Feldman helped characterize and name this phenotype. He has argued that metabolic context matters: the usual risk equations draw on populations with high triglycerides, insulin resistance, and excess weight, and LMHRs represent a different biological situation [2]. Tom Dayspring and Peter Attia hold that apoB drives atherosclerosis through cumulative arterial exposure regardless of metabolic health markers [3, 4]. Ference's causal analysis, drawn from large genetic studies, supports the view that LDL-lowering reduces risk in a log-linear, dose-dependent fashion with no apparent saturation point [1].

What KETO-CTA found

KETO-CTA was a pre-registered prospective study examining coronary plaque in LMHRs using CT angiography [5]. Its main findings are not disputed: participants developed plaque at a rate of roughly +0.8% per year, with an average increase of 18.8 mm³ in non-calcified plaque volume. LMHRs are not exempt from atherosclerosis. Every commentator on the study agrees on that.

The disputed finding was secondary: within the LMHR cohort, apoB levels did not correlate with the rate of plaque progression. Feldman and colleagues pointed to this as evidence that apoB may not be the driving variable in this population [2].

Why that finding doesn't settle the debate

Verhoeven's analysis suggests the apoB-progression result should be read cautiously for several reasons [2]. The apoB analysis was a secondary outcome in a single-arm study. It was likely underpowered to detect a within-cohort correlation. More critically, every participant had high LDL, so there was no low-apoB comparison group. Without that range, the study cannot distinguish between "apoB doesn't matter here" and "apoB matters but everyone was already so high that variation within the band was noise."

The +0.8%/yr progression rate itself is not reassuring. Verhoeven's analysis suggests this figure is comparable to what is seen in metabolically unhealthy cohorts [2]. For a group defined by supposedly excellent metabolic markers, that is a worrying result, not a green light.

The LMHR camp's most ambitious explanation holds that apoB may stop predicting risk above some threshold because arterial binding sites become saturated. That runs directly against the log-linear cumulative-exposure evidence [1]. Verhoeven, who was initially sympathetic to the LMHR hypothesis, describes this saturation argument as weak [2].

One detail to note: Verhoeven publicly retracted and reissued his analysis after critique during the review process. That kind of self-correction is what careful evidence appraisal looks like, and it bears on how his conclusions should be evaluated.

Where the evidence stands

The causal role of apoB/LDL in atherosclerosis is not in dispute. Genetic evidence, trial data, and observational studies converge on it, and KETO-CTA showed that LMHRs develop plaque [1, 5].

What remains open is whether apoB predicts rate of progression within the specific LMHR phenotype, and whether some individuals with high LDL but pristine metabolic markers progress more slowly than predicted. The single study addressing this lacked a control group and was underpowered [2]. Feldman and colleagues ran a genuine pre-registered trial rather than a thought experiment. The heterogeneity question, why some LMHRs progress faster than others, is a legitimate research question that population-level data cannot yet answer [2].

Dayspring, Attia, and Ference's causal data all point in the same direction [3, 4, 1]: high apoB in an LMHR warrants the same concern it warrants in anyone. That means imaging to establish plaque burden and a serious conversation about lowering exposure. KETO-CTA does not contradict that stance. The decisive missing piece is a matched normal-lipid control group and an adequately powered apoB-progression analysis. Follow-up papers from the KETO-CTA group are the next data to watch [5].

References

1. Ference et al. 2017, Eur Heart J — LDL causes ASCVD (EAS consensus) · PubMed 28444290
2. Physionic (Verhoeven) — LMHR: an exception to the cholesterol rule? (revised)
3. Attia #229 — understanding CVD risk, cholesterol & apoB (Dayspring-based)
4. raw/attia/attia207-antiaging-drugs-nad-metformin-rapamycin.txt
5. Soto-Mota/Budoff 2025, JACC Adv (KETO-CTA) — LMHR plaque