Of all the reasons people try a ketogenic diet, blood sugar is one of the best studied. Type 2 diabetes is, at its heart, a problem with how the body handles carbohydrate, so it makes intuitive sense that eating far less of it would help. The good news is that this is not just intuition. There is a sizeable body of trial evidence, and in 2025 researchers pulled much of it together in one place. Here is what that evidence actually supports, along with the practical questions people ask most: how fast it works, whether it can push diabetes into remission, what to eat, and the one safety point you cannot skip.
The study, in brief
A team led by Yan and colleagues published what is called an umbrella review in the journal Diabetology and Metabolic Syndrome. Rather than running a new trial, they gathered 21 existing meta-analyses, each of which had itself combined the results of many randomised controlled trials, and looked at what the weight of evidence says about low-carbohydrate, very-low-carbohydrate and ketogenic diets for blood sugar control in people with type 2 diabetes. Stacking the reviews up like this gives a broad view across thousands of participants, which is about as close to a settled picture as nutrition research offers.
What they found on blood sugar
The headline measure for long-term blood sugar is HbA1c, which reflects your average glucose over the previous few months. Sixteen of the 21 reviews found that cutting carbohydrate produced significant reductions in HbA1c, with the other four finding no significant change. That is a clear majority pointing the same way. People also tended to see improvements in body weight alongside the better glucose readings, which is part of how the effect works, along with lower fasting glucose and, in several reviews, reduced need for medication.
How quickly does it work?
Faster than most people expect. Because carbohydrate is the nutrient that raises blood glucose most, cutting it produces a change you can often see on a home meter within days. Fasting readings and post-meal spikes usually start settling in the first week or two. HbA1c is slower by design, since it averages several weeks of glucose, so the fuller picture shows up on a blood test at around the three-month mark. If you monitor at home, this early feedback is one of the genuinely motivating parts of the switch, and it is worth pairing with the macros that keep you in range rather than guessing.
Can keto reverse type 2 diabetes?
This is the question everyone really wants answered, and it deserves a careful reply rather than a slogan. Doctors now talk about remission, meaning blood sugar back in the normal range without glucose-lowering medication, rather than cure, because the underlying tendency can return if old habits do. Remission is a real, documented outcome, and it is most likely when carbohydrate restriction is paired with meaningful weight loss and started reasonably early in the disease, before insulin-producing capacity has faded too far. It is not guaranteed, it is more achievable for some people than others, and it typically holds only for as long as the way of eating does. So keto is a credible route towards remission for the right person, not a switch that flips diabetes off permanently.
The honest part: time matters
Here is the finding that an enthusiastic blog would skip and that you deserve to know. The improvements were strongest in the short term, over roughly three to six months, and they tended to fade with time. In some studies the advantage over a normal diet had largely disappeared by the two-year mark. This is not a reason to dismiss the approach, but it reframes it. A low-carb or ketogenic diet looks like a genuinely effective tool for getting blood sugar down, especially early on, rather than a permanent fix you do once and forget.
The likeliest reason is not that the biology stops working but that the diet gets harder to stick to over years. The benefit depends on keeping carbohydrate low, and life has a way of eroding strict eating plans. That points to the real challenge being sustainability rather than effectiveness, which is worth knowing before you start, because it tells you where to put your effort.
Why it works
The logic is straightforward. Carbohydrate raises blood glucose more than fat or protein, so eating very little of it means smaller glucose rises to manage, less demand on whatever insulin response a person still has, and over time often some weight loss, which itself improves insulin sensitivity. Our beginners guide explains how the diet shifts the body towards burning fat for fuel, and it is that same shift that eases the carbohydrate load the body has to cope with day after day.
What to eat in practice
The everyday menu is simpler than the science. Build meals from protein, natural fats and low-starch vegetables: eggs, fish, poultry, meat, cheese, olive oil, butter, avocado, and plenty of leafy greens, courgette, cauliflower and the like. Cut the fast carbohydrate that drives the spikes hardest, meaning bread, pasta, rice, potatoes, sugar and sweet drinks. Where a food has a label, judge it on the carbohydrate that actually affects you, which is why it pays to understand net carbs rather than react to the total on the front of the packet. Most people aim for somewhere under 20 to 50 grams of carbohydrate a day depending on how strict they are and how their readings respond.
The safety part you cannot skip
The most important point is one of safety rather than diet design. If you take medication for type 2 diabetes, particularly insulin or sulfonylureas, cutting carbohydrate sharply can lower your blood sugar enough to cause a hypo if the medication is not adjusted to match. This is genuinely something to manage with the doctor or team who prescribe for you, before and during any big change in how you eat, not afterwards. People do reduce or come off medication under supervision, but the supervision is the point. Two related cautions: type 2 is a different situation from type 1 diabetes, where the risks and rules differ, and keeping your electrolytes in order in the early weeks makes the diet far easier to stay on, which, given that adherence is the weak link, matters more than it might seem.
The bottom line
A 2025 synthesis of 21 reviews found that low-carbohydrate and ketogenic diets significantly improve blood sugar in type 2 diabetes for most people, with the clearest gains in the first few months and weaker evidence that the advantage lasts for years. It works quickly on daily readings, can drive remission for some when combined with weight loss and started early, and lives or dies on whether you can keep it up. Used with medical support and an eye on medication, it is a credible tool rather than a miracle.
This is general information about the ketogenic diet, not medical advice. If you have type 2 diabetes, especially if you take any glucose-lowering medication, do not change your diet without talking to your doctor or diabetes team first, because your medication may need adjusting to avoid dangerously low blood sugar.
Source: Yan Y, Asemani S, Jamilian P, Yang C. The efficacy of low-carbohydrate diets on glycemic control in type 2 diabetes: a comprehensive overview of meta-analyses of controlled clinical trials. Diabetology and Metabolic Syndrome. 2025;17:341. Read it here.