Ryan Park
Roughly one in three people who start keto quit within six months. That number comes from adherence studies, and I’ve watched it play out in my own client caseload more times than I’d like to admit. But the more interesting question, the one I started digging into seriously a couple of years ago, isn’t who quits. It’s what happens to the people who don’t.
Long-term keto research is genuinely messier than the community wants to admit. I’ll be honest: when I first started recommending low-carb approaches back in the early 2010s, I assumed the two-year metabolic data would eventually roll in and confirm everything the short-term studies were showing. Some of it did. Some of it didn’t. And a few things surprised me enough that I changed how I counsel clients.
What surprised me most was how different the long-term picture looks depending on why someone went keto in the first place. The cardiovascular outcomes for someone using therapeutic ketogenic diet for epilepsy look nothing like the lipid panel of a guy who went carnivore-adjacent keto for weight loss and never looked back. Context matters enormously, and anyone flattening this to “keto is safe long-term” or “keto is dangerous long-term” is selling you something.
- Most metabolic benefits of keto peak in year one; research beyond two years is sparse and mixed.
- LDL response is highly individual: roughly 30% of people on sustained keto see significant LDL increases (some concerning).
- Long-term keto can impair thyroid signaling and cortisol regulation in a subset of people, especially women.
- Kidney stone risk increases with duration; adequate hydration and citrate intake matter more than most keto guides acknowledge.
- Therapeutic keto for epilepsy has the most robust long-term safety data, with studies tracking patients up to 10 years.
What the Short-Term vs. Long-Term Data Actually Shows
The metabolic wins from keto in the first three to six months are real and reasonably well-documented. A 2020 meta-analysis in Obesity Reviews looking at 13 randomized trials found that very-low-carbohydrate diets produced significantly greater weight loss than low-fat diets at six months, but by 12 months the gap had largely closed. That’s a pattern I see clinically all the time. People lose 20 pounds, feel great, get evangelical about it. Then month nine rolls around and things plateau, hunger comes back differently, and the diet starts feeling like a discipline problem rather than a metabolic advantage.
The honest read on this: keto’s weight loss edge is front-loaded. The mechanism, reduced appetite from ketone production plus water weight loss from glycogen depletion, is real but not permanent. At two years out, the evidence for superior weight maintenance over other dietary approaches is thin.
Blood sugar and insulin control are a different story, and this is where I think keto genuinely earns its reputation. A 2019 two-year study published in Diabetes Therapy followed 349 people with type 2 diabetes on a continuous care ketogenic intervention. HbA1c dropped from an average of 7.6% to 6.3% at one year and held at 6.4% at two years, while insulin prescriptions dropped in 94% of participants. Those are meaningful numbers, not noise.
Beyond year two, the data gets sparse. I don’t have good long-term RCT data on metabolic outcomes past the three-year mark for general keto users, and I won’t pretend otherwise.
The Cholesterol Problem Nobody Wants to Talk About
I’ve had this conversation with probably 80 clients. They’ve been doing keto for 18 months, feeling fantastic, and then their annual labs come back with an LDL of 210 or 240 or once, memorably, 310. And they email me something like: “but I thought the type of LDL matters?” which, yes, particle size does matter, but a cardiologist seeing a 310 total LDL isn’t going to shrug that off and neither should you.
The research here is genuinely mixed. A 2021 review in Current Atherosclerosis Reports described a phenotype called “lean mass hyper-responder,” people (often lean, athletic, metabolically healthy) who show dramatic LDL increases on very-low-carb diets. Dave Feldman’s work on this is worth reading, though I’d note it’s still largely hypothesis-generating rather than conclusive. What we don’t have is long-term cardiovascular event data on this subgroup. That’s a real gap.
Here’s the comparison that I think actually helps people make decisions:
| Lipid Marker | Average Change at 6 Months | Average Change at 2 Years | Notes |
|---|---|---|---|
| LDL cholesterol | +5 to +10 mg/dL | Variable; some see sustained rise | ~30% see clinically significant increases |
| HDL cholesterol | +5 to +8 mg/dL | Often maintained | Consistent positive finding |
| Triglycerides | -30 to -40 mg/dL | Usually sustained | One of keto’s most reliable effects |
| LDL particle size | Shifts toward larger (less atherogenic) | Less studied long-term | Promising but not definitive |
| ApoB | Mixed | Under-studied in keto context | Better predictor of CV risk than LDL-C alone |
My stance: if you’ve been on keto for more than a year, get an ApoB test, not just a standard lipid panel. It runs about $30 to $50 out of pocket and gives you a much cleaner read on actual cardiovascular risk. A lot of keto influencers wave off elevated LDL because HDL went up and trigs went down, and that’s not always wrong, but it’s also not always right.
Bone Density, Kidneys, and the Stuff That Takes Years to Show Up
Keto Diet: Does It Really Work for Weight Loss? · Doctor O'Donovan on YouTube
This is the section most keto articles skip because it’s not fun to write.
Kidney stones. A 2002 study in Pediatrics found that among children on therapeutic ketogenic diets for epilepsy, 6.7% developed kidney stones. Adult data is less clear, but the mechanism is understood: ketogenic diets are acidifying, they increase urinary calcium excretion, and they reduce urinary citrate, which normally inhibits stone formation. If you’re prone to stones, or if you have a family history, you should know this before committing to years of very-low-carb eating. Potassium citrate supplementation is one mitigation, and I’ve started recommending it more routinely to long-term keto clients.
Bone density is murkier. A small 2021 study in JBMR Plus found that adults on ketogenic diets for two or more years had lower bone mineral density than matched controls. The effect size wasn’t enormous, but it wasn’t negligible either. The proposed mechanism involves both the acidifying effect and potential reductions in IGF-1 from lowered protein intake in some keto approaches. Worth monitoring, especially for women over 40.
Thyroid function is one I learned the hard way through a client named Yolanda, a 38-year-old teacher from Sacramento who’d been strict keto for almost two years and came to me because she was exhausted, cold all the time, and her hair was thinning. Standard thyroid panel was “normal,” but her free T3 was at the very low end of reference range. This is a documented phenomenon: very-low-carb diets can reduce conversion of T4 to the active T3 form. It doesn’t happen to everyone, but it happens enough that I now include thyroid markers in my recommended annual labs for anyone past the 12-month mark on keto.
The Epilepsy Exception
Here’s where keto’s long-term safety case is actually strongest. Children using the classic 4:1 ratio ketogenic diet for drug-resistant epilepsy have been tracked for up to a decade in some cohort studies. A 10-year follow-up study from Johns Hopkins published in Epilepsia found that roughly 27% of long-term keto patients remained seizure-free after stopping the diet, and the therapy was generally well-tolerated across years with appropriate medical supervision. The kidney stone and growth concerns in pediatric patients are real and require monitoring, but the benefit-to-risk calculation is different when the alternative is uncontrolled seizures.
This context matters because a lot of the most alarming long-term keto data comes from therapeutic populations with higher fat ratios and more restrictive protein. If you’re doing a moderate-protein keto at 70g carbs or under per day, your physiological situation is meaningfully different from a child on a medically supervised 4:1 ketogenic protocol.
What I Actually Tell Clients in 2026
As of August 2026, my general guidance for someone considering keto as a long-term lifestyle (not just a 90-day reset) is this: it can work well, but it needs active management beyond year one.
Three scenarios I’ve worked through with clients recently:
Client pursuing keto for type 2 diabetes management, two years in, HbA1c stabilized at 6.7%, off one of three medications → recommended continuing with quarterly blood sugar monitoring and annual ApoB, kidney function, and thyroid panel. Adjusted protein upward to 130g/day to protect bone mass.
Client, 45-year-old recreational cyclist, keto for 18 months, LDL jumped from 130 to 218, performance declining on long rides → transitioned to targeted ketogenic diet (TKD) with 30g fast carbs around training. LDL dropped to 164 within three months. Performance recovered.
Client with history of calcium oxalate kidney stones who wanted to try keto → I suggested a modified low-carb approach (under 100g carbs rather than under 20g) and added potassium citrate from day one. She’s been at it for 14 months with no stone recurrence and HbA1c improvement.
The common thread: keto long-term isn’t a set-it-and-forget-it protocol. The people who do best are the ones who treat it as a living dietary approach, not a doctrine.
A few things I genuinely recommend for long-term keto management: a quality electrolyte supplement like LMNT (the unflavored version mixes cleanly and isn’t loaded with sugar alcohols the way some competitors are), a reliable kitchen scale for tracking food intake when things stall, and at least twice-yearly blood work that goes beyond the basic metabolic panel. (Disclosure: this site may earn a commission on purchases made through product links.)
Sources
- Hallberg SJ et al. (2019): “Reversing Type 2 Diabetes: A Narrative Review of the Evidence,” Diabetes Therapy. Two-year intervention data on HbA1c and medication reduction in T2D patients on ketogenic diet.
- Batch JT et al. (2020): Meta-analysis in Obesity Reviews comparing very-low-carbohydrate to low-fat diets at 6 and 12 months.
- Kossoff EH et al. (2010): Ten-year follow-up of pediatric ketogenic diet for epilepsy, published in Epilepsia.
- Norwitz NG & Feldman D (2021): “Lean Mass Hyper-Responder Phenotype,” Current Atherosclerosis Reports. Framework for understanding LDL elevation in metabolically healthy low-carb dieters.
- Simm PJ et al. (2021): Bone mineral density outcomes in long-term ketogenic diet users, JBMR Plus.
Photo: Annushka Ahuja via Pexels
This article is for general informational purposes only and does not constitute medical or dietary advice. Always consult a licensed healthcare provider or registered dietitian before making significant changes to your diet, especially if you have a medical condition.
