#5: The Ketogenic LCHF Diet for T1D, 2012

In the previous post, I described my short experience (5 weeks) with The Paleo Diet (TPD) and how it affected my blood glucose (BG) and insulin doses in the quest to improve the treatment of my type 1 diabetes (T1D). The BG improved slightly, but I experienced a significant reduction in mealtime insulin doses by eliminating sugar, grains, and dairy and increasing a variety of fruits including those higher in sugar content. As I mentioned, my research was continuing and I discovered the work of Richard K. Bernstein, MD who had found that a very low carbohydrate diet improved his BG to close to normal levels and this reversed his serious diabetic complications.

I decided on February 8, 2012 to continue TPD but to eliminate fruit and fruit juice which was all that was needed for me to decrease my carbohydrate intake to about 50 grams/day and thus shift my metabolism to primarily a fat and ketone burning state: called nutritional ketosis. If you look into Dr. Bernstein’s method, he suggests 30 grams of carbs spread out over three meals a day. For me, that would limit the amount of vegetables I could eat and I thought my level of exercise would allow for a more liberal carbohydrate intake. The transition to my ketogenic low carb high fat (KLCHF) Paleo diet was fairly uneventful although I lowered the fruit gradually so that I would have to time to adjust the insulin doses.

After about 2 weeks, I had eliminated all the fruit in my diet, so now my diet consisted of primarily whole foods, cooked at home, including meat, fish, non-starchy vegetables, and tree nuts, but found I needed to add some additional fat to replace the calories in the fruit that I had omitted. I chose butter (technically not on a Paleo Diet) and in 2013 added coconut oil along with choosing fattier cuts of meat and later adding organ meats, mainly beef liver, heart, and kidney for the extra nutrients these organs contain.

I did experience a worsening of my occasional dizziness on standing after starting the KLCHF diet. After reading The Art and Science of Low Carbohydrate Living by Drs. Stephen Phinney and Jeff Volek, I realized I was not adding salt to my food nor drinking enough water. This small correction solved the problem of dizziness on standing. However, I do continue the same additional salt and water consumption to this day. In the chart below, you can see the results of adopting the KLCHF diet both during the first three months, 2-5/2012 = Feb. – May 2012, and for the rest of 2012, 6-12/2012 = June – Dec. 2012.

Keith Runyan, MD Results 2007-2012

You see during the first 3 months on the KLCHF diet, my mean BG improved from 145 to 123 mg/dl and insulin doses decreased modestly, this time both basal (from 22.2 to 20.3 IU/day) and mealtime (from 16.0 to 14.9 IU/day) insulin doses. In addition, the frequency of “acceptable” BG values in the range of 51-120 mg/dl increased from 45% to 53% and the frequency of high BG values >200 mg/dl decreased from 20% to 7%.

However, to my pleasant surprise I stopped experiencing symptoms of hypoglycemia! If you’ve never experienced hypoglycemic symptoms (sweating, fear, hunger, confusion, racing pounding heartbeat, blurred or double vision, etc.) you may not be able to appreciate how special that is. If you have T1D you will certainly have experienced it. I was and continue to be fascinated by this phenomenon and have contacted multiple physicians who do research in diabetes and hypoglycemia to see if they were familiar with this effect. I won’t mention their names as I have not asked them for permission to quote them, however, I’ll paraphrase their responses. All of them were clearly not interested in researching this observation. First, there is very little interest at the National Institute of Health (NIH) in funding research on ketogenic diets. Second, pharmaceutical and food companies would not be interested either unless there was something sell, but a KLCHF diet can be easy formulated from whole foods purchased at any grocery store. Most of the researchers thought it was an interesting observation, but doubted that blood ketone levels could be high enough to prevent the symptoms of hypoglycemia by following the KLCHF diet. Another was down right concerned that I was experiencing hypoglycemia unawareness. This is a well described phenomenon where persons with T1D who have frequent hypoglycemic episodes have either no symptoms or a reduction in the number and/or severity of hypoglycemic symptoms during hypoglycemia. Thus, the concern is when hypoglycemia occurs in persons with T1D who have hypoglycemia unawareness that they will not take action to treat the hypoglycemia and that if it occurs particularly while sleeping or driving that they could die. This is a real possibility as it is estimated that 6-10% of persons with T1D die from hypoglycemia.

In my situation however, you can see in the chart above that I had only eleven (281*4% = 11) BG values <51 mg/dl during the first 3 months or about 1 per week (11 per 12 weeks = 1 per week). Most diabetes authorities consider a BG value <70 mg/dl to be consistent with hypoglycemia, and in data not shown above, 11% of my BG values were <70 mg/dl. That calculates to 2.5 BG values <70 mg/dl per week (281*11% = 31 and 31 per 12 weeks = 2.5 per week). However, I use the <50 mg/dl figure since I have rarely had any symptoms of hypoglycemia when the BG is > 50 mg/dl since 1998 when I started on insulin probably due to the fact that my BG has always been reasonably controlled. Thus having only one BG <50 mg/dl per week during the first 3 months on the KLCHF diet, I felt would not be enough to result in hypoglycemia unawareness, although it is possible my BG was low at other times between measurements. If we use the 2.5 BG values <70 mg/dl per week instead, then hypoglycemia unawareness is a valid, but not exclusive, explanation for lack of symptoms. Thus, I will never be able to prove to myself or anyone else for that matter that I haven’t experienced hypoglycemia unawareness or that I won’t in the future. And I would never want anyone reading this to think that I condone hypoglycemia. I am just relating my personal observations of my efforts to control my BG. My position on glycemic control in T1D is that BG should be as close to normal (83 mg/dl or 4.6 mmol/L) as is safely possible by taking steps to minimize hypoglycemia.

I will be writing a future blog post on hypoglycemia and including some very interesting research studies that when extrapolated to nutritional ketosis suggest it may be possible to obtain partial protection from hypoglycemia by being in nutritional ketosis. This topic really needs to be researched in a formal laboratory setting since as I mentioned above, 10% of persons with T1D are dying from hypoglycemia.

From June 2012 to the end of the year I did experience two episodes of symptomatic hypoglycemia. These episodes were different than those I had experienced before. Yes I had sweating and didn’t feel right, but since Feb. 8, 2012 my cognition during hypoglycemia seems to be markedly less impaired at worst and outright normal the vast majority of the time since starting the KLCHF diet. I could be wrong, but I suspect this is due to the brain fuels, acetoacetate and beta-hydroxybutyrate, that our livers make when dietary carbohydrates are restricted. Again, I will never be able to prove this is not due to hypoglycemia unawareness, that is why research needs to be done.

Other changes from June 2012 to the end of the year included a further reduction in my mean BG from 123 to 107 mg/dl and  my HbA1c decreased from 7.1 to 5.6%. The frequency of “acceptable” BG values in the range of 51-120 mg/dl increased from 53% to 62% and the frequency of high BG values in the range 121 to 200 mg/dl decreased from 36% to 26%. This was accompanied by an increase in the frequency of BG values <51 mg/dl from 4% to 6%, two of which were accompanied by symptoms as mentioned above. My basal insulin dose also increased from 20.3 to 27.1 IU/day which I feel was due to my attempt to lower BG values throughout the year (a lower BG requires more insulin all else being equal).

As you may recall from a previous post, one of the reasons I was seeking out a better method to control my T1D was my plans to do The Great Floridian Triathlon in October of 2012. It turns out that a KLCHF diet may provide some fuel flow advantages for endurance exercise as discussed in The Art and Science of Low Carbohydrate Performance. Because of this, my plan was to take a source of sugar with me at all times and measure my BG during the bike and run portions of the race and only eat the sugar if my BG started to fall. Up to the race during training I had been taking sugar to prevent hypoglycemia and this generally, but not always, lead to hyperglycemia. However, I noticed in the unusual circumstances of hypoglycemia while exercising that I had two consistent symptoms, namely a slowing of pace and numbness of skin usually in the pelvic area (don’t know why just there). I never experienced confusion or an unawareness of hypoglycemia while exercising even prior to starting the KLCHF diet. So I was not terribly concerned about developing hypoglycemia and not being aware of it during The Great Floridian Triathlon. I had made the decision not to take my morning insulin with breakfast since if I had it would have been peaking during the 2.4 mile swim. I was especially hoping to avoid hypoglycemia during the swim since it is difficult to check BG and consume small amounts of glucose while swimming. During the race, my BG was actually a little higher than I anticipated and would have liked (150 – 250 mg/dl). But all in all, I was successful in avoiding hypoglycemia throughout the race and completed the ironman distance triathlon in 15.5 hrs without any problems. I did not need to eat anything, no sugar, no food, just water. And I finished with a smile.

Keith Runyan 2012 GFT finishline cropped

Finishing The Great Floridian Triathlon, Oct. 2012.

In the figure below, I graphed the one week moving average for both the BG and total daily insulin dose for the year 2012. The moving average smoothes out the data and shows the up and down movements more clearly.

2012 BG and Insulin One Week Moving Average.png

In the next blog post, I’ll review my continued efforts to improve my BG control in the year 2013, the addition of coconut oil to my diet, and blood ketone measurements.




  1. Pingback: #19: February 2016 Update on My T1D Management | Ketogenic Diabetic Athlete
  2. fulmar2

    Keith – I just recently discovered your blog, and have been devouring the information you provide. I like your writing because you provide so much data – and also admit when there is something you don’t know. A lot of material online present opinion as fact, which isn’t always accurate. I was diagnosed T1D 65 days ago at the age of 39, but adopted the LCHF diet right away (thanks to reading online, not to my doc (who didn’t take a stand on the diet issue)). It took a while to titrate my dose, but I can maintain normal blood sugars with just Lantus (~12U, maybe now titrating down). I’m very interested in your experience with hypoglycemia because today I was on a 100 mile bike ride (7 hours), and measured 5 times – each reading right around 65 +/- 3. In other a few other instances, I have dropped below 60 (all 3 times were when micro-doses of alcohol were combined with exercise). In those 3 situations, I don’t really start feeling hypo until a reading of <50mg/dL, and even then the symptoms are very minor (but enough to make me want to check). My theory on this is that perhaps years of endurance training (sometimes running out of food) may have caused an adaptation to lower glucose levels. I'll never know, though, because there is no way to obtain data from those pre-diabetic days. I do worry about the safety of running around so low, but don't feel a strong urge to eat. Perhaps this can serve as another data point of athletes on LCHF. I am especially curious to see what the non-diabetic blood glucose profiles of endurance athletes are like on the LCHF diet… and trying to emulate that… Too bad that you'll almost never see a non-diabetic with a CGM!


    • Keith Runyan, MD

      Thanks for your comment, Brian. Just a few comments regarding your observations. Hypoglycemia unawareness is more likely to occur during sleep, during exercise, after exercise, and after consuming alcohol. Avoiding alcohol would help reduce it and I think there really aren’t any health benefits of alcohol. So I avoid alcohol and don’t recommend that others with T1DM consume it. When hypoglycemia does occur in T1DM on a ketogenic diet, it should be treated with glucose (tablets or liquid) if there are symptoms or if no symptoms could also be treated with food, but not to excess. There is no advantage to staying hypoglycemic: it should be treated whether or not one is hungry or having symptoms or not. The reason for not having symptoms of hypoglycemia while on a ketogenic diet is likely a combination of brain utilization of ketones and hypoglycemia unawareness, in my opinion. In addition to sleep, exercise, and alcohol, antecedent hypoglycemia is a common reason for hypoglycemia unawareness: another reason hypoglycemia should be treated. There are more details on hypoglycemia unawareness in blog posts #12 and #29.