Keto works. For most people who follow it strictly, it produces meaningful weight loss in the first 3–6 months. The harder question — the one the research is much less enthusiastic about — is what happens after that. This post covers what the long-term adherence data actually shows, the biological trade-offs that appear over time, and what to do instead if you're looking for something that holds beyond a year.

What Keto Is (And Why It Works Short-Term)

The ketogenic diet restricts carbohydrates to roughly 20–50g per day — about the amount in one medium apple — forcing the body to shift its primary fuel source from glucose to ketone bodies derived from fat. This metabolic state, ketosis, produces several short-term effects that drive weight loss: appetite suppression (ketones have a direct effect on hunger hormones), reduced insulin levels, and the loss of glycogen-bound water weight in the first 1–2 weeks.

The short-term results are real. A 2025 narrative review in Current Nutrition Reports confirmed that ketogenic diet interventions consistently produce reductions in body weight, BMI, waist circumference, and fat mass in obesity studies — with most trials running 8–24 weeks.

The issue is not whether keto works. It's whether it's a mechanism that humans can and will sustain for years — and whether doing so is without cost.

The Adherence Problem: What the Data Shows

Adherence data is where the keto narrative gets uncomfortable.

StatPearls' 2024 clinical review of ketogenic diets — one of the most comprehensive clinical summaries available — reported adherence rates in therapeutic keto programs (for epilepsy, where motivation is highest) of roughly 70% in year one, dropping to 38% by year three. In weight loss populations, where the stakes are lower, dropout rates are documented as even higher.

The reasons people stop are consistent across studies: social events become difficult, family meals require separate preparation, food choices at restaurants are severely limited, and the constant vigilance required for carbohydrate counting becomes psychologically exhausting over time.

Keto is also a diet that doesn't forgive interruptions well. Unlike a Mediterranean-style eating pattern where occasional deviation doesn't break the mechanism, exiting ketosis — even briefly — requires a multi-day return period. Every holiday, birthday, or work dinner that involves a carbohydrate becomes a reset event. For many people, the cognitive and social cost of maintaining this over years is higher than the benefit.

This is not a willpower argument. It's a design argument. A dietary approach that requires near-perfect adherence to function is structurally less sustainable than one that works within normal human social behavior.

The Long-Term Biological Trade-Offs

Beyond adherence, there are biological considerations that emerge over longer time periods.

LDL Cholesterol

The cardiovascular lipid response to keto is the most clinically significant variable for long-term safety. Most people on keto see triglycerides fall and HDL rise — both favorable. But LDL response is heterogeneous: a meaningful subset of people experience substantial LDL increases, including elevated levels of small dense LDL particles that carry higher cardiovascular risk. This response is not predictable from baseline and requires monitoring. Without regular lipid panels, long-term keto carries an unquantified cardiovascular risk for some individuals.

Muscle Mass

The 2025 narrative review noted minor but consistent decreases in lean body mass and skeletal muscle mass on ketogenic diets — particularly without resistance training. This matters more as people age, when maintaining muscle mass is increasingly important for metabolism, mobility, and long-term health outcomes. Protein intake on keto is often adequate but the combination of calorie restriction and low carbohydrate availability does put muscle at somewhat greater risk than higher-carbohydrate approaches.

Animal Research Signals

A 2025 study published in Science Advances — updated from a 2024 preprint — found that long-term ketogenic diet in mice produced hyperlipidemia, hepatic steatosis (fatty liver), and impaired glucose tolerance from disrupted insulin secretion. Animal studies don't translate directly to humans and the dietary conditions were extreme, but this research flags biological mechanisms worth monitoring in humans doing long-term strict keto.

Gut Microbiome

Fiber intake on keto is typically low — high-fiber foods (legumes, whole grains, most fruit) are largely excluded. Fiber is the primary substrate for gut microbiome diversity. Preliminary human research suggests keto produces measurable changes in microbiome composition, with some studies showing reductions in beneficial bacteria. Long-term consequences are not yet established, but this is a legitimate area of ongoing investigation.

Who Keto Actually Works For Long-Term

Being honest about this matters. Keto is not equally appropriate or inappropriate for everyone.

Strongest evidence base:

• Epilepsy management — particularly drug-resistant pediatric epilepsy, where it has decades of clinical use

• Type 2 diabetes management — several studies show meaningful HbA1c reduction and potential for diabetes remission under medical supervision

• Short-term metabolic reset for people with significant insulin resistance

Who tends to do well long-term: People who genuinely don't miss carbohydrate-heavy foods, have flexible social eating contexts, enjoy high-fat foods, and find the appetite suppression effect outweighs the dietary restriction. These people exist — they're a minority, but they're real.

Who tends to struggle: People in social eating cultures (shared meals, restaurant-heavy lifestyles, food as social ritual), people who experience significant cognitive load from tracking, athletes who need carbohydrates for performance, and anyone who finds the restriction psychologically amplifies food preoccupation rather than reducing it.

The Weight Regain Reality

Even for people who succeed on keto short-term, the weight regain data follows the same pattern as other restrictive diets. More than 95% of people who lose weight through calorie restriction regain it within 3–5 years — and keto weight loss is primarily a calorie restriction mechanism, even if it doesn't feel like one.

When people stop keto, glycogen stores refill, water weight returns, and appetite regulation — which was suppressed by ketones — normalizes. Without the behavioral habits that would maintain a lower intake in the absence of ketosis, weight returns. This is not a keto-specific failure. It's the same pattern that affects all restriction-based approaches.

The biological mechanisms of weight regain — hormonal compensation, adaptive thermogenesis, adipose tissue memory — operate regardless of which diet produced the initial loss.

What Works Instead

The alternative to asking "which diet should I follow?" is asking "what eating behaviors can I sustain indefinitely?" Those are different questions with different answers.

The research on long-term weight maintenance consistently identifies behavioral predictors — consistent meal timing, adequate protein, high food volume, low dietary restraint — not specific dietary patterns. The Harvard Plate structure captures these principles without requiring carbohydrate elimination: half vegetables, quarter protein, quarter whole grains.

This structure naturally reduces caloric density, maintains satiety through protein and volume, and doesn't require monitoring a macro target. It works at restaurants, at family dinners, at work lunches. That social compatibility is not a minor consideration — it's one of the primary reasons some approaches hold and others don't.

If the reason you tried keto was to lose weight without counting calories, there are approaches that achieve that without the elimination requirement. If it was to reduce sugar cravings, addressing eating habits and food structure produces the same outcome with a less restrictive mechanism.

"Keto works, and I've seen clients lose real weight on it. What I've also seen is most of them gain it back when life gets in the way — a holiday, a stressful month, a partner who doesn't eat keto. The diet didn't teach them anything about their relationship with food. It just changed what they were allowed to eat."

— Irene Astaficheva, PN1, PN-SSR, GGS-1

Honest Limitations of This Analysis

Keto has legitimate therapeutic applications that this post is not dismissing. For drug-resistant epilepsy, it's one of the most effective non-pharmacological interventions available. For type 2 diabetes management under medical supervision, the evidence is meaningful. These are not the contexts being discussed here — this post addresses keto as a self-directed long-term weight management strategy for otherwise healthy adults.

The long-term human safety data on strict keto remains genuinely limited. Most studies run 6–24 weeks. The animal research signals (fatty liver, glucose intolerance) are concerning but not conclusive for humans. This is an honest gap in the evidence — long-term strict keto in humans is under-studied, not definitively safe or unsafe.

FAQ

Does keto work for weight loss? Yes, short-term — the evidence is consistent for 3–24 week interventions. The weight loss is real. The sustainability question is separate from the efficacy question.

Why do people regain weight after stopping keto? Several mechanisms: glycogen-bound water returns immediately, appetite suppression from ketones disappears, and the behavioral habits that would maintain lower intake in the absence of ketosis are often not in place. The biological compensation mechanisms that drive weight regain after any calorie restriction also apply.

Is keto safe long-term? The honest answer is: we don't fully know, because long-term human data is limited. The signals that exist — lipid variability, minor muscle loss, microbiome changes — are reasons for monitoring rather than immediate alarm. Anyone doing keto long-term should have regular lipid panels and ideally medical supervision.

What's the difference between keto and low-carb? Keto specifically requires ketosis — typically under 50g carbs/day — to be effective. Low-carb is broader and can range from 50–150g per day. Low-carb diets allow more flexibility, more food variety, and easier social eating while still reducing carbohydrate intake meaningfully. Many people who find strict keto unsustainable do better with a moderate low-carb approach.

If keto isn't sustainable, what should I do instead? Focus on eating behaviors rather than dietary rules: adequate protein at each meal, high vegetable volume, reduction in ultra-processed food, and awareness of hunger and fullness signals. These changes produce a meaningful calorie reduction without requiring macro tracking or food elimination — which makes them significantly more sustainable over years.

Bottom Line

Keto is not a scam. For the right person, in the right context, with medical monitoring, it can produce meaningful and maintained results. For most people using it as a self-directed weight loss strategy, the adherence data tells the same story as every other restrictive diet: it works until it doesn't, and when it stops working, the weight comes back.

The more durable path is building eating habits that don't require a metabolic state to maintain — habits that work at dinner parties, on holiday, and in every other real-life eating context that a ketogenic diet struggles to accommodate.

Start Here

The Eated Habit Wheel identifies the single eating behavior with the most leverage for your specific patterns — no elimination required. Free, 5 minutes.

The Eated app is free to download on iOS — built around habits, not macros.