My month with the newest kid on the block & a hypothesis on the #Fiasp effects.

Fiasp First two weeks AGP

To some level I can’t quite believe that I’ve been enjoying the company of Fiasp for a whole month (well, in February, four weeks constitutes a month…). At first it was a joy, with massively reduced bolusing time and huge efficiencies, however as the month has progressed, our friendship has soured somewhat. I’ve been needing more and more of it and it’s not been much fun trying to work out what’s been going on.

But let’s take a step back and look at the stats from the flip over. First up, we need a comparison, so here is the data from mid-February to mid-March. We have a months worth running from 17th Feb to 21st March, giving 31 days, with a limits set at 4 being low and 10 being high, in accordance with similar studies.

In the grand scheme of things, these aren’t too bad looking sets of numbers, so comparing this to Fiasp is quite interesting. For the month of 24th March to 21st April, the data is shown below (and let’s remember this includes Easter and a little too much chocolate):

Statistically, from an Hba1C perspective, there doesn’t seem to be a huge amount of difference (the benefits of a Hybrid Closed Loop, I guess), but the AGP graph appears to show that afternoon and late evening highs as a result of eating have been reduced, overall resulting in a slightly greater time in range.

Now this can be split into two different periods. The start of the month using Fiasp and the end of it. The start is the top, the end the bottom.

Fiasp First two weeks AGP
Fiasp last two weeks AGP

What you can see in the second is that the 25th-75th percentile range seems wider and that post prandially, I appear to be marginally higher.

Fiasp First two weeks stats
Fiasp last two weeks stats

We see that, if anything, the second two weeks saw a greater amount of time low (almost twice as much) as the first two week period.

Putting the two together sets of data together, this is the start of where we start to see some of the unexpected behaviour kicking in.

Over the first two weeks, the average TDD was 46.9u, and in the second, 52.7u. This is a 12.3% increase in TDD, that could be explained by Easter, but that seems unlikely to be the only cause. I’ve discussed in other posts how I’ve seen what I felt was random behaviour and unexpected requirements for large amounts of insulin, and as a result, I’ve probably rage bolused a little too much, resulting in the increase in lows.

Since I’ve mentioned this, I’ve heard from others that they’ve seen similar behaviour, with one person saying they’d “bolused with it and it was like using water” and another mentioning that they’d found themselves using 1.5x as much insulin and coming out in welts, so were going back to Novorapid.

As is my wont, I decided to go and dig into Fiasp to try and understand what might be going on. This stuff is just insulin, right?

Well, yes and no. Firstly, my research took me into the realms of the Novo patents for Fiasp, and this particular item: Preparation comprising insulin, nicotinamide and arginine – WO2013186138A1

So what is this? it’s the 2013 filing of a patent application for the use of Nicotinamide (the water soluble version of vitamin B3) and Arginine in insulin.

Why is this interesting? It’s use of the word “Surprisingly” in the description that Nicotinamide speeds up the absorption of Aspart. And also the suggestion that the discovery that the use of Nicotinamide causes the insulin to become less stable, creating more “High Molecular Weight Proteins” (HMWP) or insulin dimers instead of monomers. The Arginine is used to reduce this affect. The issues with HMWP in insulin are detailed here.

Why would this be interesting? What’s the deal with Nicotinamide and Insulin? Why use the term “Surprisingly”? Well, further digging presents some research, and a not insignificant amount, that has shown that Vitamin B3 is linked to insulin resistance and a drop in insulin sensitivity, as detailed in the lists of research studies from Indeed, this article from Diabetes Care in 1999 looked at using Nicotinamide as a method of reducing the affects of Type 1 Diabetes and was viewed at the time by some as a cure, and happens to mention insulin sensitivity effects.

The key here is the use of vitamin B3 supplements in fairly large doses affecting insulin sensitivity. The sample sizes are fairly small, but the results don’t seem insignificant. With Fiasp, we seem to be using something that appears to be known to affect insulin sensitivity to speed it’s absorption. And although the sizes of the dose of B3 are miniscule, it still seems to be having an effect, and one that builds up.

There is a certain irony that Novo are using something that has been shown to decrease insulin sensitivity to speed up the absorption of insulin.

At the end of my first month with Fiasp, I’ve redone my insulin sensitivity testing, and the results are really quite stark. My ISF has got worse by 37.5% (1u now drops me 1.6mmol/l vs 2.8mmol/l at the start) and my IC ratios have increased accordingly. But at these levels, Fiasp is back to working as it should be. To the extent that the fiasp/katsu experiment that was previously done has been repeated, with very similar results and a flat glucose line, a couple of hours after eating, but that is a lot of insulin….

My hypothesis then is that by adding a form of B3 to the insulin in order to make it absorb more quickly, the slow build up has caused insulin sensitivity to drop over the month, and the results were the variable figures I had been seeing. In addition, where I had cannulas going in to me where there is less of me, I saw increased resistance as the B3 was unable to dissipate as effectively.

I need to spend time at this new level to see if it is maintained or gets worse again, but it seems to make sense. I also wonder if the insulin sensitivity issues that B3 can cause that seem to be one-offs are linked to this same mechanism or whether it’s simply a part of the process of insulin desensitization and giving an inappropriate dose for the food that’s been eaten?

The other thing I noted from the studies that links to is that in some cases, blood lipids were also affected, which suggests that this is something that may need to be observed carefully in all Fiasp users.

Given this, what recommendations should we take away from it?

I think I’d start by suggesting that at least once a week while getting started on Fiasp, you need to check your ISF and IC ratio. For those of us that have reported it, it has changed as we’ve used Fiasp for longer. And by changed, I mean has gotten substantially worse.

The second thing I’d question is whether people want to use Fiasp, given this potential effect on insulin sensitivity. I know that there are people that value this being low, and I suspect that using Fiasp this won’t be maintained.

Finally, it probably makes sense to have your blood lipids checked before and after using this insulin to see if there are any detrimental effects. The research on this point was somewhat inconclusive.

Whilst so far it’s only an N=3 population of people reporting changes in insulin sensitivity, it looks as though the excipient that makes Fiasp faster also, over time, makes it less effective per unit used.

I’m not sure how clever or sustainable that is in the long run.


  1. Another really interesting article Tim I’ll make sure I pay careful attention to changes in ISF if I ever get my hands on any! Interested to see if the resistence keeps changing over time.

  2. Very useful information and food for thought. I’ll have to think about whether I want to use Fiasp when it is approved in the U.S.

  3. It’s just occurred to me..if the theory on B3 plays out to the rest of the Fiasp users and it stays at the same price as Novorapid, that could be very lucrative for Novo.

    • Referring to “if the theory on B3 plays out to the rest of the Fiasp users”, do you mean that the due to the ISF dropping, users will require more Fiasp, so increasing Novo’s sales? Or have I got this wrong?

  4. Super interesting… great work Tim! Definitely cooling my lust for fiasp as an ingredient to my openAPS closed loop

  5. Hi Tim, has your decline in insulin sensitivity continued or did it stop at approx 50%? Thanks, Tim

  6. Thanks for this in-depth analysis! It’s always nice to find confirmation of something I’ve suspected. I’ve trialed Fiasp for 3 days now using multiple daily injections (neither of which are ideal conditions for solid conclusions), but have noticed a severe decrease in insulin sensitivity. I’m fairly good at carb counting and have kept my diet to standard meals (I know my normal NovoRapid dose for these meals) and my IC ratio has gone from 1:15 with NovoRapid to 1:9 with Fiasp.

  7. I thought it was just me, my ratio’s have changed dramatically for the worse, also the quick onset that I enjoyed so much in the first few months has now become non-existent and often feel like it’s a clear hour and a half before I see any shift in blood glucose levels. As a result I’m taking much more insulin than I used to.

  8. Terms like “surprisingly” are used in patents because claims are only granted for novel discoveries. Anything obvious would not be approved.

  9. I’ve also been having (suspected) problems with Fiasp – but almost the opposite.
    I’ve been using it for six months. For the first 4, there wasn’t much change in sensitivity compared to NR. However, over the last two months, my ISF has INCREASED by almost 50% while my IC has got worse.
    However, that’s not the whole story. People have talked about their BG getting stuck high. This also seems to be a feature – my ISF being inversely proportional to BG & getting stuck. I have a suspicion that my IC may not actually be changing, it’s just the effect of the ISF reducing that makes it seem as if I need more Insulin to deal with the carbs?
    Having gone from being pretty stable to decidedly unstable over two months, today, I decided to go back to NovoRapid to see if it reverses the trend.

    I wonder if the B3 / Nicotinamide sensitivity ratio is the same across the population or alternatively if the rate of absorption / sensitivity is related to the ambient BG level?

  10. Have been searching for comprehensive information on Fiasp and this fit the bill.

    This is a great article and gave me enough knowledge for my research into insulin sensitivity testing.

Leave a Reply

Your email address will not be published.