Skip to Content

Improvement in Glycemic and Lipid Profiles in T2 Diabetics with a 90-day keto Diet

ByCrossFitOctober 17, 2019

Eleven diabetic women were instructed to follow a ketogenic diet for 90 days. Dietary instructions were simple: Reduce carbohydrate intake to less than 30 grams per day while consuming ~20% of calories from protein and the remainder from fat. Ketone levels were tested weekly, with all 11 women successfully maintaining at least mild ketosis (i.e., detectable elevation of blood ketone levels, indicative of at least some adherence to the KD) for the duration of the study.

All 11 women lost weight, with an average weight loss of 9 kg. HbA1c decreased dramatically, from an average baseline of 8.9% to 5.6% at 90 days. Nine of the 11 women completed the study with an HbA1c below the cutoff for a diabetes diagnosis (6.5%). Improvements in multiple cardiovascular risk factors were also observed, including a decrease in blood pressure, an increase in HDL cholesterol (43.1 to 52.3 mg/dL), and a dramatic decrease in triglycerides (177.0 to 92.1 mg/dL). Both individual and group changes are summarized in the figures below.

In total, these results indicate a very low-carbohydrate ketogenic diet can reverse diabetes while driving substantial weight loss and reducing heart disease risk within 90 days. When compared to existing therapies, such as insulin secretagogues, this suggests the ketogenic diet is both more efficacious and has fewer negative side effects than pharmacological alternatives for diabetes treatment.


  1. Drug-induced weight gain, Initiation of sulfonylureas versus metformin is associated with higher blood pressure at one year

Comments on Improvement in Glycemic and Lipid Profiles in T2 Diabetics with a 90-day keto Diet

Comments 11

Sort By

Henrik Holm ThomsenOctober 18th, 2019 at 4:30 am

No control group....CF HQ, please stick to CrossFit and leave the science to the ones who know how to read scientific articles. You are are cherry picking, it is almost embarrassing

jr WildOctober 18th, 2019 at 6:30 am

This is a Benjamin Bikman study, which CF refers to. He has some great lectures in Youtube. You seem to imply that there is no science without a control group? This seems to be a prospective cohort, you follow them after intervention. If any drug got these results, it would be a blockbuster. In their study you'll find references to other studies, even with a control group. Since you may not read the hole study, here is their conlusion: We submit these data as modest, yet undeniable, evidence that a dietary intervention that restricts carbohydrate and emphasizes unrestricted consumption of protein and fat elicits a favorable metabolic state, including the dramatic reduction in HbA1c. Moreover, such an intervention, unlike various insulin secretagogues [24, 25], has the favorable side effect of meaningful body weight and blood pressure reductions. JR

Emilio CORBEXOctober 18th, 2019 at 7:23 am

Unfortunately I agree with Henrik. The interpretation of the results is misleading. The decrease in blood pressure, the improvement of the lipid profile and the decrease of HbA1c are the result of Keto Diet or simply caloric restriction? Losing an average weight loss of 9kg ! This weight loss could probably explain alone this impressive results, whatever the macro nutrient removed (fat, carbs, proteins). Futhermore, even if the lipid profile is improved, what do we know about the state of the liver? oxidative stress? the weight or lipid profile of the subjects one year after the intervention? It would be too bad if thousands of people reading this interpretation this morning think to see in it a "scientific revolution". In addition, the restriction of carbohydrates generally implies an extremely high fat diet ... which can lead to extremely important systemic inflammations, which greatly favors metabolic diseases in the long term (metabolic endotoxemia). In order to maximize their chances to being healthy, people should stick to the recommendations of Walter Willet (Eat, Drink and Be Healthy). This book 's recommendations are made on the best current scientific knowledge. Always good to debate! Have a good day, Emilio

Clarke ReadOctober 18th, 2019 at 5:18 pm

In the abstract, this is a valid pushback. In context, I take some issue with it. In fact, I am becoming less sympathetic to it in the abstract as well. The abstract reasons randomized controlled trials are praised, and single-arm trials are criticized, are so well documented I need not recount them here. Of course, compliance with these rules has costs. A trial with a proper control group is more complex, more expensive and will likely take longer than a single-arm design. At minimum, this means it will take longer for this evidence to emerge; in some cases, it may mean a feasible trial which can yield some valuable data will not be done at all if this bar must be crossed in all cases. This trial tests an intervention for a condition for which there is no current, effective standard of care (i.e., that consistently reverses the underlying pathology) other than surgery, in which previous research (reviewed elsewhere on the site) has shown the intervention to not be harmful. I am unaware of any interventions, pharmaceutical, lifestyle or otherwise, which lead to a similar magnitude of improvement in HbA1c alongside simultaneous improvements in lipid profile and weight, with the exception of fasting (most studies of which would be subject to similar criticism) and bariatric surgery. This is also far from the only trial using a ketogenic diet, or severe carbohydrate restriction, to improve glycemic control; these other trials have shown results of similar direction and magnitude. As such, while they do not meet the objective standard of a randomized controlled trial, these results nonetheless have value. Contextually, given your own background, you likely understand how difficult it is to receive public funding for trials testing low-carbohydrate (let alone ketogenic) dietary interventions. This pragmatic constraint limits the construction of more rigorous trials. To suggest we should dismiss, suppress or otherwise ignore this information because it does not take a particular form is, in some ways, to damn this potentially valuable research thread to remain on the outskirts of scientific thinking in perpetuity, barring large, privately funded trials. Fundamentally, we face two choices when dealing with research in any emerging scientific area, or any scientific area in which the tested intervention is at odds with standard of care. We can continue to require this bar to be reached for any information to be considered valid. Or we can interpret the results with necessary caution but without dismissing them whole-cloth. It is a balance of risk versus opportunity. Applied categorically, this does open the door for pseudoscience to intrude, and that is a valid criticism. If a critical scientific reader or writer is to add value, it is by judging which observations have merit and which do not, and to interpret the validity and value of each trial on its own merits, independent of and within the context of the broader literature. Categorical application of abstract rules has led to the incidental suppression of nascent fields of research, the glacial pace of intellectual progression, and the predictable dominance of well-funded (rather than well-supported) hypotheses that characterizes the current nutritional literature. We can do better.

jr WildOctober 18th, 2019 at 6:32 pm

This is more to Emilio, I like the reasoning of Clarke. In my acquired bias, weight loss is a result of something useful. This something is the driver of these other favorable effects. In the end of the day, however, these two have a strong correlation. Which came first, no problem with that. Energy-in can be divided in two components: Eeaten and Einside. If you can reach to your endogenous energy inventories, i.e. fat, then you need to eat less. This is what happens with carb restriction, the more the faster the bigger the effect. It is nice and easy, if hunger signal is your guide, not clock or habit, to eat. If you succeed in harvesting energy from your stores, you naturally want to eat less -> your calories-in diminish. In order to achieve this, you need to control your insulin demand-levels. This has both with carbs and timing to do... I accidentally went this road. First changed the breakfast to egg-and-something, then within months wondering, should I eat breakfast while I am not hungry? Answer no, follow your hunger. I am still a "brunch" guy, unless I seldom feel hungry in the morning. This study with metabolicly derailed persons should work similarily (with amplified effect due to damaged metabolism). Diminish the stimulus, get access to storages, and follow your hunger (eat as much protein/fat that makes you satisfied). No need to count calories-in, trust your body (biochemistry). JR

Greg GlassmanOctober 19th, 2019 at 1:27 pm

Henrik, this conversation quickly rose past your inputs. Perhaps it is you who should refrain from engaging in matters science?

Greg GlassmanOctober 19th, 2019 at 2:07 pm

Henrik, I think decency would require you to respond to Clarke. He went to considerable effort to respond to your complaint of no control group. You don't sit in greater command of science or the scientific method than everyone around here, you can be sure of that. Your arrogance doesn't square with your modest achievements.

Scott GillinOctober 22nd, 2019 at 4:39 pm

Henrik it is quite disappointing to see a physician/scientist try to discredit the dissemination of information. Earlier this year there were a number of people on this site who seemed to gravitate towards ad hominem- attacking the person rather than the argument. It seems that you are using this technique against CrossFit as an institution. If you would like to use you scientific training to analyze the study and it’s limitations, as well as what can be learned from it, to further educate the people of this website, that would be great. I am not sure if you have a clinical practice but the medical system as well as the dietary guidelines are failing. Society is getting fatter and the incidence of diabetes is rising. Medications are not the solution. They do not address the root cause, sedentism and excessive refined carbohydrate ingestion. No randomized controlled studies were done on CrossFit. Is it the intensity, variability, functional movement, community aspect? Does it matter? If it works it is a solution. Can other fitness programs work? Does this mean CrossFit does not work? It no different with nutrition. If you have a program that works for this epidemic great. Does that mean another program does not work? Does it matter whether it is the lack of carbohydrates of the reduction in calories? If reducing refined carbohydrates and sugar mean that you are not hungry all the time and thus you consume more calories, isn’t this desirable? Nutrition, health, fitness, reversing chronic disease are all part of CrossFit’s lane. As a physician I welcome any lifestyle solutions to lifestyle diseases.

Glenn HodgesOctober 18th, 2019 at 1:30 pm

I truly love the fact that this organization provides literature and a place where others with more, less, or contrary information can come together and discuss their perspectives. Thanks for the added perspectives. They help those of us who are not experts better understand the range of the possible when reading studies like this one.

Greg GlassmanOctober 22nd, 2019 at 12:08 pm

Emilio, Agree with Henrik? What on the stupid part or the fucking rude part? Since you’re weighing in it must be the stupid part. (Henrik thinks this is the wrong place to have these conversations because we aren’t smart enough.) I think Clarke addressed the issue of a lack of control for this intervention well enough that Henrik owes apology for a baseless response and being a dick at the same time. You can be wrong here, and you can be a dick. Don’t try both at the same time. “This weight loss alone could probably explain this impressive results”, you offer. We see weight gain and similar results though less often. To this we could say, “This weight gain alone could probably explain this (these) impressive results!” The empiricist in me tells me that you need a better theory. And then there’s this: “In addition, the restriction of carbohydrates generally implies an extremely high fat diet ... which can lead to extremely important systemic inflammations, which greatly favors metabolic diseases in the long term (metabolic endotoxemia). In order to maximize their chances to being healthy, people should stick to the recommendations of Walter Willet (Eat, Drink and Be Healthy). This book 's recommendations are made on the best current scientific knowledge.” I’m not sure quite where to start, but I can share with you that it’s not getting the attention it deserves because of the greater value to focusing on Henrik’s misunderstanding (thank you Clarke) and contention that we shouldn’t even be talking about these kinds of things in a place like this. Everyone is smarter here than to believe that anything published in peer review is going to be a “scientific revolution” on the subject of what we should eat, but…were thousands stupid enough to think that such a thing were likely the net result would be an unprecedented wave of health that would expose Walter Willet as still irrelevant. (He’s not been useful to the cause of reversing trends in chronic disease for decades. I cannot imagine having to read his book.)

Shakha GillinOctober 22nd, 2019 at 4:01 pm

I am a practicing physician. This study is significant to me, control arm or not. What this study tells me, is that when I have a patient with T2D, they may benefit from a carb restricted, keto diet as presented in this paper. My patient might go from being a diabetic to a non diabetic, improve their lipid profile, and their LFTs. And when an L1 trainer reads this, or anyone else who is a T2D, or has a loved one with T2D, or has a client with T2D, this can be a solution for that person. This paper has major clinical relevance. (Oh, and I still have the option of meds and surgery if I want to wait for that control arm).