Type 1 Keto

Reviewing Diabetes Control on a Keto Diet


Here is the CGM trace of the two-week project. This is a true real-world trace. The project was not done in the ‘perfect’ environment. For  various reasons, there was scope to screw-up these weeks. An increase in sedentary behaviour and a couple of stress days with grazing. And poor sleep. You can see the effect by looking at days three and four, the 4th and 5th of June, in the table above. I was grazing on keto food, but it made a difference to my control. The high plateaus are insulin resistance due to snacking. I decided not to inject insulin because I  know from experience that this trace for me represents insulin resistance.  The only way to get that down is to do physical activity. I understand that this pattern is not due to insulin deficiency. If it were, the glucose would keep rising. But it never does. It just plateaus. It’s as if the body has other priorities and parks the glucose at a safe enough level while it gets on with whatever it has to do. Then it sorts out the glucose. I don’t know.  I don’t think that is the case for everyone. Nor for me all of the time. It was the stress that caused the rise in glucose.  

I had a reduction in sleep during the first week, especially.  I need around seven and a half hours optimally but was down to six hours for several days. For two nights, it was less than that. Lack of attention to these so-called pillars of health take their toll after a while. Sleep, physical activity, and stress all play their part in glucose control. But diet plays by far the most significant role.  Stress, physical activity, and sleep ratchet up the level of difficulty required to achieve tight glucose control. On some days, there were swings as I tried to stabilise the glucose levels by injection of insulin (one or two units maximum), food, or physical activity. I used to get similar problems with a higher carb diet, but the swings then were far more exaggerated. I would get glucose highs into the twenties mmol/l on occasion.  Then the glucose would crash down to hypo levels. Now that was a roller coaster.   

Nowadays, the norm for me is to have long periods where the control is smooth, where no management is needed. I can get on with life.  They are the days to aim for. And they are not rare.  Pottering about all day. It can be done in an office environment as well if the manager is sympathetic. But I can’t ever imagine reasonable control on a high carb diet.  

The green band highlights the range 3-7. Time In Range is good in this example. Time in Range is turning out to be an important marker for complications. It can only be measured on a CGM.  

I am fortunate to be able to self-fund a continuous glucose meter(CGM). It is undoubtedly the single most crucial device after insulin to improve diabetes control. Everyone I know who has one feels the same. Every clinician off the record agrees. It is so frustrating to watch the slow process of negotiation around CGM. People are suffering daily. Let’s hope things change as soon as possible.  

I have published real-time blood glucose for the two weeks of the project.  I use insulin pens to manage my diabetes.  I use half the insulin now compared to the bad old days. The good thing about a CGM is that I can micromanage the glucose. Occasionally I needed to take glucose. But it is mercifully rare nowadays compared to the carbing-up regimen I was recommended to follow for the first 20 years of diabetes.   

  I  managed an hour of physical activity every day. On six of the days, I did a continuous hour of running. You can spot it on the glucose traces at around 18:00. I did not need to worry about managing my diabetes, and all I had to do was to tie on some running shoes. Such small doses of insulin make everything gentle and safe.  My glucose got low but never into a hypo sensation.  

 You might notice that some of the traces are erratic.  These are due to the sensor playing up. It does seem to have a wobble from time to time, and this week’s experiment was used with a sensor that was a week old already. You get used to the sensor after a while. A sudden increase of 5 mmol/l and a similar drop over a quarter of an hour has no other explanation. I have used two brands and four models of those brands. I find that on occasion, they all do this. They are great but are a tool only and should be used with skill. There is around a 15-minute delay in getting the accurate glucose reading to the receiver.  It is okay if you set your alarms high enough. But for running, I follow the trend and not the actual glucose reading on the receiver. 

There are few pieces of medical equipment that have a delay of 15 minutes. Imagine an electrocardiogram in a cardiac ward reading 15 minutes behind!  But having said that, I am not complaining. CGM is transformative in Type 1 management. Infinitely better than finger-prick tests. I am so lucky to have one. Finger prick tests are not always accurate, either.