#21: March 2016 Update on My T1D Management

This month was marked by both some whacky blood glucose (BG) readings and successes including two new personal bests in weightlifting, a 185 lb. clean & jerk and 140 lb. snatch. I also had my first episode of symptomatic hypoglycemia in 2016 which I over treated and ended up with a BG of 260 mg/dl. In addition, I had two high morning readings over 200 mg/dl which are quite unusual that I think were related to my changing exercise pattern i.e. having to take 10 days off from weightlifting due to a lacerated thumb. My BG readings during the 10 day reprieve from weightlifting were excellent which made March not so bad in the glycemic control department considering the whacky BG readings. I did daily aerobic exercise (walking and cycling) while my thumb was healing, but the stopping and starting weightlifting and weightlifting in general definitely is not T1D friendly as far as BG is concerned. But I am OK with that since I enjoy weightlifting very much. I am optimist that eventually I will adapt to it and my BG will too. For example, my BG prior to and after my 2.5 hour weightlifting session including the 140 lb. snatch were 106 and 98 mg/dl, respectively on March 31.

Below are my BG reading along with exercise type and time so you can see how the exercise affects my glycemic control.

Post 21 Blood Glucose - Mar. 2016

The table below shows the summary of current and previous BG results along with insulin dose totals.

Post 21 Mean BG

My total insulin doses increased due to an increase in basal (Lantus) insulin which I thought might help lower my fasting BG and maybe mitigate the increase in BG while weightlifting. This month I increased my protein intake by a half pound of meat at dinner daily to hopefully add some muscle mass over time which is a good thing as one ages, but also might improve my weightlifting results. This did not result in a significant increase in dinner mealtime insulin (Humalog). The average daily mealtime insulin dose at breakfast was 6.2 IU, dinner dose was 2.8 IU (compared to 2.6 IU in February), and the average non-meal correction dose was 1.9 IU. I think the higher breakfast dose (compared to dinner) is related to the dawn phenomenon combined with a relatively lower dinner dose due to the increased insulin sensitivity related to afternoon exercise.

The table below shows the summary of current and previous BG variability results. Some results were improved, some not, compared with the previous month.

Post 21 BG Variability

The actual daily insulin doses and totals are shown in the graphs below.

Post 21 Insulin Dose

My breath ketones since June 2015 are shown below. I remain in continuous nutritional ketosis.

Post 21 Ketonix Breath Ketones

The graphs below show a new metric I conceived of this month. On the x-axis is the breakfast (AM) and dinner (PM) mealtime insulin dose (Humalog) plotted against the change in BG i.e. Pre meal BG minus Post meal BG on the y-axis in mg/dl. Thus positive values represent a reduction in BG and negative values represent an increase in BG after the meal. Also, note that I removed 3 of 31 AM points and 9 of 31 PM points from the graphs where the Post meal BG was not in the range of 51-120 mg/dl. The rationale for this is to eliminate mealtime doses that were “incorrect” so to speak. This way the graph shows both that larger doses of insulin reduce BG more (obvious), but more importantly even when the resulting Post BG was in an acceptable range, that there is a wide variation in the amount of BG reduction for any given dose of insulin. This is the frustrating part of T1D that we just have to live with.

Post 21 Selected Meal Pre-Post BG

The graph below is exactly the same but shows the non-meal Humalog correction doses that resulted in a Post BG in the range of 51-120 mg/dl. This graph includes 11 of the 15 Humalog correction doses given in March. Again, the main feature is the wide variation in the response to any given dose of insulin.

Post 21 Selected Other Pre-Post BG

Also, note the R squared values (correlation coefficients) in the three graphs above. The closer the R squared value is to 1, the better the fit to the linear regression line. In an ideal world the reduction in BG should be related to the dose of insulin given assuming the meal consumed is constant from day to day. I do place a great emphasis on keeping meals constant from day to day in terms of amount of food and macronutrient composition, so I don’t think that explains the wide variation observed. In the case of Humalog correction doses (last graph), there is no meal to confound the results, yet the response is even more varied.

My next blog post will cover prevention and treatment of insulin resistance, hyperinsulinemia, prediabetes, and type 2 diabetes.

Till then ….



  1. Svet Pavlovsky

    Interesting! It seems like your PM Humalog variation range (~90) is even higher than the AM variation range (~60). It looks challenging 🙂 then I realized to master the BG from the variation graphs. I have not looked on the BG values like that but rather done the average over month.


    • Keith Runyan, MD

      Svet, I probably didn’t explain the graphs well enough. The y-axis (vertical) is the difference between the pre-meal BG and the post-meal BG in the first two graphs for breakfast (AM) and dinner (PM), respectively. It shows whether my BG went up or down and by how much depending on the dose of mealtime insulin (Humalog) given. So there are no actual BG values in these graphs, just differences. For example, in the first graph (AM) you see a BG difference of about 90 mg/dl after a dose of 7.5 IU of Humalog. That point represents the difference between my pre-breakfast BG value of 139 and my post-breakfast BG value of 51 which equals 88 (mg/dl). So 7.5 IU of Humalog taken with breakfast resulted in my BG dropping from 139 to 51 mg/dl on that particular day. I’m going to look at other time periods and see if there might be a better way to predict the dose of mealtime insulin. My current method is an educated guess based on the previous few days responses. This method might help others as well so if it works for me, I’ll write a specific blog post on it.

      Liked by 1 person

  2. Svet Pavlovsky

    Hi Keith, thank you for your detail explanation. I did understand the Y axis. What I meant is that your differences in the evening have a range of 90 for a certain insulin dose, while in the morning for a certain dose your difference range is around 60. For PM for 3.5 IU you can get differences 120 or 30 so the range is 90. For 2.5 IU you can get 70 or -20 so the range is 90. For AM 6 IU the values are between 30 and -30 so the range is 60. I meant that your uncertainty or variation in BG difference is higher in the evening. Maybe it’s related to your excise regime in the evening. Hope we understand each other.


  3. Warrick

    If one begins using insulin, is there no way of tapering off it on lchf diet? The dramatic results reported by Bernstein and others on reducing insulin injection but not to zero puzzle me. Although T1D myself (per antibody, fasting glucose and hba1c) I resisted going onto insulin. I cut carbs in my diet to manage. So far for 18 months and quarterly Hba1c now in healthy range. So far so good. Any idea how long lived the antibodies are? I’m interested to check but the test is expensive. I recently purchased a fasting insulin test and came back well within normal range.


    • Keith Runyan, MD

      Warrick, you are to be congratulated for finding and implementing the LCHF diet for T1D. It is likely that the diet has contributed to your remaining off insulin after 18 months. This is called the honeymoon period. Blog post #18 was on this topic. I have yet to see a permanent cure of T1D from an LCHF diet or by any other means which is why Dr. Bernstein and I do not make any claims of such. Autoantibodies are used to distinguish between T1D and other forms of diabetes which are primarily due to insulin resistance, T2D or gestational diabetes for example. The autoantibody levels decrease with time and the inability to detect them does not necessarily mean the beta-cell destruction has stopped. As pointed out in blog post #18, maintaining blood glucose near normal will help preserve the remaining beta-cells which is important because the more beta-cells you have the better the blood glucose will be even if/when you start insulin in the future. I would be interested in hearing how you do in the future on the LCHF diet which I obviously think you should continue.


      • Warrick

        Thanks Keith – I’m certainly hoping the honeymoon lasts a long time then. However, I had always thought this referred to the period after beginning exogenous insulin when there is a reduced requirement for a period compared to the initial dose. I’ve never been on insulin, metformin and for a short period glipizide.
        My reason for wondering about coming off insulin is because there is a relatively recent paper indicating even very long term T1D with insulin actually produce clinically relevant amounts of endogenous insulin. 2015 Most People With Long-Duration Type
        1 Diabetes in a Large Population-Based Study Are Insulin Microsecretors http://dx.doi.org/10.2337/dc14-0871
        I’ve also recently noted some suggestion that gluten free assists preserving beta cell mass. Of course going without carbs from grains would also make for being gluten free – potentially a win either way! http://dx.doi.org/10.1136%2Fbcr.02.2012.5878 as an example.


      • Keith Runyan, MD

        The honeymoon period does refer to a reduction or cessation of insulin after starting insulin for T1D. The insulin therapy resolves glucotoxicity and allows the beta-cells to improve insulin secretion. However, I think it is appropriate to use the term to refer to an improvement in glycemic control in T1D after some other therapy like LCHF, gluten-free diet, or Paleo diet http://www.ijcasereportsandimages.com/archive/2015/012-2015-ijcri/CR-10582-12-2015-toth/ijcri-1058212201582-toth-full-text.php Regardless of the therapy, your blood glucose should be close to normal (83 mg/dl or 4.6 mmol/l). If your blood glucose is elevated and not corrected, the beta-cell function will deteriorate with time and blood glucose will continue to increase. Glipizide should be avoided in both T1D and T2D as it forces already impaired beta-cells to secrete more insulin regardless of the blood glucose level and worsens beta-cell function. Finally, being a micro secretor is good for T1D glycemic control but the insulin secretion is not enough to control blood glucose without exogenous insulin.