Practiceby Lab 10 Get started ›
Crossfire

My LDL cholesterol went up on a ketogenic diet. Is that a problem?

What the experts say

Stephen D. PhinneyMD, PhD · physician-scientist; co-founder and former Chief Medical Officer, Virta Health

Not on its own. Get particle number and ApoB, and read the rise against the whole risk panel before you act on it.

What it rests onIn his company's one-year trial (Bhanpuri et al. 2018, reported in the Virta post) calculated LDL cholesterol rose slightly while total LDL particle number and ApoB were unchanged and the distribution shifted toward larger particles, alongside a large fall in triglycerides, a rise in HDL and a ten-year ASCVD risk score change of -11.9 percent. His mechanical point is that standard-panel LDL is calculated rather than measured, and a large change in triglycerides skews the calculation, so on this diet the calculated number is least trustworthy exactly when it moves. The evidence is his own company's trial: Virta Health sponsored it, it tested Virta's programme, and he is a founder and employee offered stock options.

Source: virtahealth.com ↗
Christopher D. GardnerPhD · nutrition scientist, Stanford Prevention Research Center

Treat it as a real cost of the diet to weigh against its benefits. Get your lipids checked and expect LDL to rise.

What it rests onThe Keto-Med randomized crossover trial he led: LDL cholesterol rose about 10 percent on the ketogenic arm and fell about 5 percent on the Mediterranean-plus arm, while triglycerides fell 16 percent against 5 percent. He attributes the pattern partly to what the ketogenic arm removes, since dropping legumes, fruit and whole intact grains lowered fibre and several nutrient intakes. Read the whole abstract before characterising him: HbA1c did not differ between the two diets and improved from baseline on both, and he grants the triglyceride advantage. He is not arguing the ketogenic diet failed; he is arguing it carries 'potential untoward risks from elevated LDL cholesterol and lower nutrient intakes' and was 'less sustainable'. No commercial relationship found.

Source: pubmed.ncbi.nlm.nih.gov ↗

Overview

Phinney's data come from adults with type 2 diabetes inside a supervised programme holding sustained ketosis for a year. Gardner's come from adults with prediabetes or type 2 diabetes doing 12 weeks per arm, which is short enough that Phinney's own adaptation argument applies to it, and he would say so. Neither studied a healthy adult trying keto for weight loss, which is who is most likely to be reading this.

Where they agree

Both agree calculated LDL rises on a ketogenic diet in a meaningful fraction of people. Both agree triglycerides fall substantially. Both agree that someone doing this should have their lipids measured rather than assume. The dispute is entirely about what to do when the calculated LDL number rises and the particle measures do not follow it.

What would settle it

Outcome data, not panel data: a trial with cardiovascular events rather than lipid fractions as its endpoint, in people sustaining nutritional ketosis for years. Failing that, ApoB and particle number reported in a trial not funded by a company selling the diet would settle how much of the divergence is the biology.

Practice does not pick a winner here

You do, and your call is dated, kept, and yours to change. Follow the people you already trust and what they publish turns into a short daily practice.

Get started

Free to sign up and use. We do not ask for a card.