Dr Chris Huff

I can't quote tweet @nicknorwitz because I'm blocked, so you'll need to view my screenshots for context. Nick mentioned upfront that there's been "confusion" around the data. No one is confused. What you're seeing are valid scientific and ethical concerns around selective reporting, interpretation of risk, and the way this research is being framed publicly. Dismissing this pushback as "confusion" is just evasive nonsense to prevent having to engage with it. (1) Yes, the primary outcome (NCPV) was shown in Figure 1. But the actual numbers were completely absent from the text and had to be shared post-hoc on X. That’s not standard practice, especially when the median increase was 18.8 mm³, which is more than 2.5x higher than what your protocol predicted. Omitting this result is not being transparent and leads to misunderstanding of the study’s core finding. (2) That most LMHRs were not rapid plaque progressors depends entirely on how “rapid” is defined. When LMHRs who are lean, normotensive, insulin-sensitive, and inflammation-free show greater plaque progression than patients with diabetes, hypertension, and metabolic syndrome, that should be a wake-up call. The issue isn’t whether everyone progressed equally. The issue is the average progression was worse than in people with far more traditional risk factors. (3) Higher relative progression from baseline is a good point in theory, but the problem isn’t about percent change. The absolute change in NCPV (18.8 mm³) puts LMHRs on par with people who have type 2 diabetes in other studies. Even among participants with zero baseline CAC, PAV progressed similarly to those at intermediate ASCVD risk. That’s what makes the findings so concerning. (4) Heterogeneity is real, but it doesn’t negate group-level trends. Smoking and diabetes also show heterogeneity in who gets sick and how fast. That doesn’t stop us from calling them risk factors. The same standard should apply here. If LMHRs, on average, show accelerated progression of unstable plaque, then the burden is on us to treat that seriously, even if some individuals seem protected. (5) The lack of association with apoB is a classic case of a compressed exposure range. You can’t detect a dose-response when every participant already has very high ApoB. That’s like studying only 2–4 pack/day smokers and finding no correlation between packs smoked and CVD. Without a control group with low ApoB, this result can’t be used to argue that ApoB isn’t a driver of progression. (6) We all agree the LMHR phenotype is scientifically fascinating and absolutely worth studying. But when we promote a group with extreme lipid elevations as “metabolically healthy” and potentially low-risk, while they’re showing faster plaque accumulation than people with known disease, we have an ethical obligation to present that risk clearly. This is especially true when the messaging influences people’s medical decisions. Stay curious guys, just not too curious in case the results don't go the way you wanted.

If these are points that @BudoffMd raised, that's concerning and unexpected. He’s a respected leader in cardiovascular imaging, and I wouldn't have expected this level of hand-waving around the results. (1) "Crucial data wasn’t hidden.” Okay, technically, NCPV was included, but only as a figure, without commentary, numbers, or context. It was the primary endpoint, yet not even mentioned in the results or discussion. Only after public pressure did the lead author share that the median NCPV increase was 18.8 mm³, which was more than 2.5x greater than the protocol’s original prediction. Just take this one on the chin. You guys fucked up. Was this done on purpose? I don't know, but it's a very real issue in how key results were communicated. (2) “The study doesn’t show keto causes plaque progression.” No one credible is claiming causation. That’s a straw man. The concern is simple: the LMHRs in this study, despite being metabolically healthy, experienced plaque growth on par with or worse than people with diabetes, hypertension, and metabolic syndrome. That’s alarming. Whether the cause was keto, sky-high LDL, or something else, the outcome should raise red flags, not be deflected with semantics. (3) “They focused on PAV instead of NCPV due to CTA literature precedent.” This was not a neutral decision. NCPV is often regarded as the more relevant markers of vulnerable plaque. It also happened to be the one that showed a larger effect size. When a pre-registered primary outcome produces unexpected results, downplaying it in favor of more favorable secondary endpoints is a textbook example of selective reporting. That it was peer-reviewed doesn’t absolve it (and the peer-review clearly failed here). (4) “Relative % changes are misleading due to low baseline PAV.” The baseline PAV was indeed low (~1.6%), so percentage change alone isn’t enough. But the absolute change of 0.8% overall and 2.4% in those with CAC >100 is still worse than comparator groups with active disease. This isn’t about being “misled” by math. It’s about what the numbers mean. The fact that such a “healthy” group progressed so much, so fast, is cause for concern, not minimization. (5) “These were treatment-naive participants, unlike other studies.” Exactly! And that makes the findings even more troubling. If people with ideal health markers and no pharmacologic interventions are progressing faster than those with multiple comorbidities on medication, then the LMHR phenotype cannot be assumed to be low-risk. That’s not a flaw in the comparison, it’s a major insight. It suggests that something (most likely the extremely elevated LDL and ApoB) is overriding the expected protection of good metabolic health. (6) “Some individuals were high-risk (e.g., CAC >100), and that explains the outliers.” Sure, but again, this was supposed to be the healthiest, most ideal subset of LMHRs. Many others were excluded for not meeting strict criteria. If this is the best-case scenario, then average or below-average LMHRs likely face even higher risk. We can’t use subgroup variation to hand-wave away the average result. And let’s not forget that the reported values were medians, not means, so any extreme outliers wouldn’t have inflated the results. The elevated progression we’re seeing reflects a central tendency, not a few high-risk individuals dragging the average up. (7) “The science isn’t settled.” Of course it isn’t. That’s precisely why we must interpret early data responsibly, especially when it challenges popular beliefs. But when findings suggest accelerated plaque growth in a group often told they’re safe, it’s ethically irresponsible to minimize, reframe, or delay open communication of that risk.

I just finished writing a blog post about the Keto-CTA study for anyone interested. The LMHR phenotype does not seem protective against atherosclerosis. On the contrary, it appears to be associated with greater plaque progression than we’d expect. alexleaf.com/lean-mass-hyper…