The three steps of managing insulin

the three steps to keto

Step 1

Ensure that the basal dose of insulin is accurate. This is the insulin that is injected once or twice a day and is unrelated to mealtimes. On pumps, there is a way to programme a basal insulin infusion rate. Basal insulin is the insulin that is injected to counter the rise of glucose that is generated, not through the diet, but by the background metabolic processes that raise glucose. Insulin works against glucagon. If insulin is low, then glucagon will be high. If glucagon is high, then glucose is high. If insulin is high, glucagon is low.  If glucagon is low, then glucose is low.

If there is an imbalance of insulin and glucagon, then the glucose level will be high or low. We need to use the basal insulin in the correct dose to get in balance with the glucagon. Glucagon is active whether or not the person is eating. It has interactions with other hormones such as cortisol and adrenaline. So basal insulin is essential for this reason.

Step 2 

Revise how to count carbs and read the labels on food packets. There is so much sleight of hand played by food advertisers that it is difficult to assess some food products. There is a module to guide you through advertising speak in the free section of the site. .Counting carbs on a real food diet is easier but is still required.

Step 3

Rapid-acting ( bolus) insulin estimation. This is done by counting the carbohydrate and proteins in the meal and estimating the insulin dose to cover the meal. It is the same method that is taught at diagnosis. The main difference is that with small amounts of carbohydrates, there will be low doses of insulin needed. Also, there is a need to allow for protein in the meal on a keto diet.

Sound familiar? It’s the fundamentals of all Type 1 management. The only difference is the type of food will be ketogenic. If people are using a pump, the basics are the same.

Let’s go through this in more detail, one step at a time. Before doing so it will be necessary to ensure that your patient has access to a blood ketone meter. The calibration phase is a good time to introduce the ketone meter so that ketones can be monitored once or twice a day during this phase. This will help the patient to feel confident in ketone measurement and they will also begin to see ketone levels rising as ketones are metabolised. This is also a good exercise for the clinician to do. It will also enable the patient to discover their ‘normal’ ketone level on their chosen diet. It is, perhaps surprisingly, not that high. In a survey of 120 people on a keto diet with Type 1 on, 75% had ketone levels between 0.5 and 1.5. However, some have personal normals higher than this up to 3 or more mmol/l. The key thing is to establish a patient personal normal so that they are more likely to identify potential DKA if they are in a situation where they are at risk.

Blood ketone meters are available on prescription in the UK. Every person with Type 1 is potentially at risk of DKA whatever diet they choose, so there are no exceptions to prescribing.

 Step 1Basal Insulin Calibration. This will differ depending on whether patients are using a pump or injections. Mixed insulin ( Mixtard and all those types, often with a number in the name) will not suit a low-carbohydrate lifestyle. People on mixed insulin will need to convert to a basal-bolus regimen. Calculating the relative amounts of basal and bolus in mixed insulin is easy as the insulin is specified by the mixture, typically 70:30. There might be a good reason for recommending a specific formula, so a lower carbohydrate diet might be more suitable for this patient.

Those on both a pump or basal-bolus regimens will need to assess their basal insulin requirements. The principle is to measure glucose during a fasted state. They will then be sure that the basal insulin is just covering the glucagon and other ‘background processes’ and not food.

The best and safest time to start is upon waking. Miss breakfast and measure blood glucose. Ideally, with a continuous glucose meter or hourly by fingerprick if one is unavailable. Keep a pump at its usual basal rate. For pens, inject the patient’s usual basal dose. Many people with Type 1 inject basal insulin twice a day. Start with what they are already using and refine it later. The dose of basal insulin that most people are taking is probably close to the optimum, as it will have been perfected over months or years by trial and error. The optimal basal dose should be calculated independently of food. So, patients can see quite clearly that by injecting the regular dose of basal insulin and then not eating for a while, their blood glucose levels will indicate if the basal dose is optimum. The test can only be done when fasting. Skipping breakfast is usually the most comfortable option and aim for a late breakfast. Test blood glucose as often as possible during this period. Ideally, as mentioned, a continuous glucose meter (CGM) will be available. 

What you are looking for is blood glucose that is stable at around 4-6mmol/l. A rise or fall in glucose levels throughout the basal testing will indicate if the patient is taking too much basal, in which case the glucose will trend downwards towards the low end. It might be that they will need to take glucose if a hypo happens, and they should be made aware of this. If this happens, abandon the testing at this time and redo it again another day, possibly with a reduced basal dose. A suggestion is to repeat the test with a two-unit reduction of basal insulin each time it is required.

Repeat the process with a late afternoon and overnight fast to get a complete 24-hour profile. If blood glucose rises, it means an insufficient dose of basal insulin. On repeating the test, increase the basal dose by a unit or two at a time until a stable trace is achieved. If the blood glucose is reducing, it means that the basal dose is too high. Adjust the dose of basal insulin, estimate the required correction and repeat the fast. If there is a flat trace, the dose is perfect.

In this image, the basal dose is too high. Reduce by two units and repeat the calibration. Correct the hypoglycaemia with glucose tablets. 

In the image, below, the basal dose is just right

A perfect basal profile

You will then be confident that the basal insulin dose is accurate. It might need a revision later, but you can be confident for now that the basal dose is set up. One thing to remember is that some people can get a spike in blood glucose in the early hours or after waking. This is the Dawn Phenomenon or Dawn Effect. It is due to various hormones working against insulin. Usually, growth hormone, adrenaline, and cortisol can cause this. Growth hormone is a hormone that pulses throughout the night and antagonises insulin action. So, the blood glucose can be raised on waking.  It is more common in younger people, especially during growth phases.  Adrenaline and cortisol are stress hormones that typically add to the Dawn Effect. It is not uncommon for patients to find that their glucose rises on rising in the morning even before they have had any food or drink. That is likely to be caused by stress hormones. Dawn Effect is best managed by a tiny dose of a unit or two of rapid-acting bolus insulin. The basal dose is unlikely to be a cause of this. Typically, basal insulin doses are associated with shallow glucose gradients, and bolus insulin doses and Dawn Effect have steeper gradients.

When you are basal testing, cover the full 24-hour period. Continual fasting might be asking too much initially. Two periods of 12 hours might be more manageable. Each time the exercise is attempted, you must ensure that patients have taken no rapid insulin in the previous 5 hours. Any rapid-acting will potentially confuse the results. Some of the modern ultra-rapid insulins like Fiasp might not last 5 hours. In practice, it is closer to three hours. But 5 hours guarantees that the rapid insulin will not be active. And individuals respond, individually. It is an art.

 Always do basal tests in conditions that are entirely typical for the individual. Don’t complicate basal testing with unaccustomed strenuous exercise or unaccustomed stress. Infection or a change in the usual daily routine will equally not produce accurate results. Females might notice a change in basal requirements before and during the menstrual cycle and initial testing in a non-menstrual phase is recommended. 

There are a small number of people who, despite their best efforts, are unable to achieve a glucose profile that is stable throughout a 24-hour period. Even splitting the basal dose may not make enough difference. For these people, it might be necessary to consider using a pump to change basal delivery rates even hourly throughout the day. Everyone’s diabetes is unique to them, and it is likely that, for these few people, the mechanisms of glucose control are too erratic to rely on the sort of profiles that pens can deliver.

Once the basal dose has been calibrated correctly, steps 2 and 3 are easy.

The final steps of counting carbs and calculating bolus insulin doses are far more manageable when people don’t have to worry about the carbohydrate in the meal. Small amounts of carbohydrates mean small amounts of insulin, which in turn means more safety. Because of a very low-carbohydrate diet, the total daily insulin requirements will reduce. It will be mostly the rapid-acting insulin that is reduced because of the reduction in carbs. As a very rough guide, most people approximately halve their total daily insulin requirements on a ketogenic diet. If they also have significant insulin resistance and manage to resolve this, it will be a greater reduction. It is individual, and there are no targets.

If you prefer to be more cautious, for example, if the patient finds it more acceptable, or if you are managing complications ( a module in the Advanced Keto Course covers this), you may want to try reducing carbohydrate intake one meal at a time, say breakfast for one week. Then add the evening meal or lunch. With a pump, the principles are the same.  Optimise the basal insulin and bolus the rapid-acting insulin to cover carbohydrates and protein at mealtimes.  

If you decide on to a ketogenic amount of carbohydrates, the effect of protein on glucose control will become apparent. Protein will cause a plateau around 2-3 hours after meals because of the way in which protein is metabolised. There will more likely be a plateau rather than a spike, and it might be better to adjust the timing of injection or pump bolus to manage this effect. A basic rule for estimating the protein bolus is half the bolus dose of insulin compared to the dose for the same number of grams as carbohydrates. For example, for 20g of carbs and your ICR is ten, 2 units of bolus insulin are required. For 20g of protein, it is 1 unit of insulin. In the same way that insulin doses change for carbohydrates, the same applies to protein. Be prepared to adjust as needed in order to achieve normoglycemi