Site Changes

I am using the 780 and I will change my site and my basal bloodsugars will seem okay according to my cgm but as soon as I eat my bloodsugars will spike. My carb factor is good. My question is, how would I determine if my site is working properly when eating?

1 Like

1: I have to give my bolus 20-30 minutes before I eat. That can be challenging, especially when eating out.

(Also, good luck estimating carbs from a fancy restaurant meal!!)

2: physical activity seems to influence both insulin absorption and insulin effectiveness. So if I’m sitting around, the recently bolused insulin seems to take its time as opposed to when I’m moving.

3: I have experienced times where the basal insulin keeps me level but the bolus seems not to “work” but I don’t know if that is a site issue for sure, there isn’t any reliable reproducibility of the phenomenon. Except perhaps #2, above.


I just want to add to what @MBW wrote, I am convinced that for some of us there are no hard and fast rules when it comes to glucose regulation, there are just so many variables.

I think that in those using a pump, because the basal drip is so small it may be absorbed quicker than bolus which is usually much larger.

This may be better with the T:Slim because the infusion is fairly slow, a few tenths of units at a time. Not knowing if that is true with other pumps.

Just spitballing here.

1 Like

I’ve experienced the same thing: a site change seems to go fine but then the first meal bolus is really ineffective. This isn’t the case 100% of the time, but it’s frequent. I don’t consider these bad sites because they work fine later but they are sites that for some reason need to be “broken in”
I use a t-slim and steel infusion sets.
I’m sorry to say, I’ve never found a way to predict whether this will happen with a new site


The only time that this has happened is when my site is in blood. I will remove the site and there is a tinge of red in the cannula. Could my 780 be reducing my bolus too much because that is what it does for breakfast. I haf my carb ratio as low as 2.8 and I still ran above 10mmol.

I only test my infusion sites for accuracy when my system doesn’t seem to be behaving as expected. When I do I skip to step 3 below. Only when I have to change everything at one time, like after an MRI, do I start at step 1 below .

If possible, I try to adjust the timing of changes so I don’t need to change the infusion site and sensor within 6 hours of each other. If I do it closer, it’s difficult to know with certainty which is off or by how much. I’ve had poor sites for each, and occasionally both at once. When that happens I test the CGM first using carb tablet surges and syringe boluses.

Everyone using insulin should know how much a fixed amount of carbs raises their BG when they are “within range”, and by how much 1 unit of insulin will lower their BG, so than they can handle hypo, and so they can maintain reasonable control when their pump sensor, or both aren’t working.

I’ve tested my low glucose correction factor many times using a BGM and a glucose tablet, but it can be calculated from R and F. (That number is nonlinear for low BG values and needs to be raised 50 to 100% as BG drops lower below range. )

I= units of insulin, C = grams of carbohydrates, dBG = change in Blood glucose

If I/C= R, and I/dBG = F , then
(I/C)/( I/dBG) = R/F and
dBG/C = R/F

eg (in mg/dL) If I/C= 1:10 and I/dBG = 1:30 then dBG/1 carb gram = 3 mg/dL

or for every 4 gram glucose tablet, BG should rise 12 mg/dL

Infusions site test

  1. Change your infusion site when your BG should be in range and stable between meals.
  2. Give a correction of the lesser of 1/2 your hourly basal delivery or 1 unit. This is to counter the startup/depot effect
  3. Thirty minutes later do a finger stick test, then eat one 4g glucose tablet.
  4. Thirty minutes later do another finger stick test.

I do this with my pump in manual mode when sitting down to make things easier for myself. If I am confident that my CGM is tracking accurately then I can use it and just observe the CGM values, and observe how my pump reacts.

  1. Watch your BG over the next 30 minutes. Unless you were very active It should have risen by about R/F. (+/-10%) and then flattened or dropped depending upon your pump algorithm. If the rise was much lower then automaticslly infused insulin was countering it well.

CGM sensor change

  1. Test a CGM sensor for rough accuracy at stable BG by comparing it to a BGM,
  2. If not within +/-10%, consider calibration. Wait 30 minutes and retest.
  3. Eat a glucose tab
  4. Observe when the climb flattens.
  5. Do another finger stick. Unless the pump interfered (reacted) by doing a basal adjustment, your BG should have risen by about R/F. (+/-10%). If it’s less, the sensor may need more time to stabilize or it may need calibration at its after-meal peak. .

I think of it this way; my pump delivers insulin into my skin, it then takes hours to get to my veins at which point it gets used in 2-4 minutes. So, for me, it all comes down to insulin in the skin, waiting to get to somewhere useful (my veins).

I can see this with pump failures; sure Fiasp works in 15 minutes, but if my pump fails (like the canula comes out) I’m still not dead a couple of hours later at the point where I notice something is seriously wrong.

So a site change certainly means going from one place currently saturated with insulin to another place which is not. Basal/bolus; so what? It’s exactly the same chemical; insulin.

One approach is to delay eating for 8 hours after a site change; 8 hours is about the time insulins used in pumps take to get out of the skin. This approach sucks. So if you eat within that period of time the approach to determining whether the site is working is whether you hit 400mg/dL, rapidly. Given that you have a CGM it’s somewhat less dangerous; look at the upslope on the CGM. If in doubt whack in some insulin, not with a bolus, with a syringe/pen!

You can calculate an hour’s, or two hours’, basal and do that then stop the basal on the pump, or you can do what I do which is to wing it; I can always eat ice cream, or chug juiced oranges. Of course you shouldn’t do what I do; I’ve been a T1D since 1972 most of which time I didn’t give a damn and now, well.

1 Like

It doesn’t take 8 hours for rapid or fast insulin, pumped or injected, to get through the intersticial layer of your skin into the capillaries of your circulatory system.

That can’t happen even if the site is heavily scarred beneath the skin surface. No cell including scar tissue can survive without constant contact with interstitial fluid and/or capillaries to supply oxygen and nutrition. Dead cells decompose and get removed.

You are mistaken, exaggerating, or negligent about being dead a few hours later if you didn’t notice that your cannula came out.

There is a well documented depot effect where multiple small infusions of insulin over an area are absorbed more quickly than one large one, but that effect is over a short period of time. The other “depot” effect is for relatively small quantities of insulin, 5 units or less, and temporary.
A significant startup delay in a new site is most likely to be caused by a cannula restriction scarring or a capillary being punctured,. Slight bleeding under the skin near the tip of the canula and clotting happens frequently and goes unnoticed most most of the time.

Such a clot will eventually dissipate but a person who is paying close attention to a CGM might not be satisfied with the speed that resolves A person doing MDI is less likely to notice it because of the velocity of injections forces its way past the clotting better than pump infusion.

I know that Tandem is aware of new site problems because their set change procedure has a specific alert to check blood glucose levels 2 hours after a site change. I don’t know if any other pump maker does something similar, but it would seem likely.

Every new site, pump, injection or sensor is a potential problem. I think of pumped insulin as being like water, our bodies as crop that needs it, and us as farmers irrigating our fields.

Each site has a slightly different “geography”. No matter where insuiin is delivered, it will flow downhill. Even if it’s trapped on the hardest material, there will be cracks. If enough insulin is delivered onto impermable ground, it will overflow and find its way into the ground. As farmers we need to learn the best places to have the insulin delivered by trial and observation.



What I am saying is that it takes many hours; up to 7 for Fiasp (from the package insert) for all of the insulin to get to the main circulatory system (veins, arteries). Every insulin we use has a “pharmacodynamics” section which gives this information.

Insulet advise replacing the Omnipod within an hour of the failure/expiry of pod. This is because even with fast acting insulin the peak effect of the insulin is over an hour after injection; 90 minutes for Fiasp. The actual peak serum concentration occurs somewhat earlier - about 63 minutes for Fiasp with the first detectable insulin around 2.5 minutes after injection. This is in the “pharmacokinetics” section of the package insert.

The single number that is most helpful is the “half life”, the time for the effect (etc) of the insulin to reduce by 50%. Quoting the Fiasp insert, “The apparent terminal half-life after subcutaneous injection of FIASP is about 1.1 hours.” Using that number:

Time (hours) Effect
1 100%
2 53%
3 28%
4 15%
5 8%
6 4%
7 2%
8 1%

It doesn’t matter when the insulin was originally injected except that the first 60 minutes (for Fiasp) are spent raising the effect. After the first 60 minutes (or so) the effect of the insulin falls off as in the table. So if everything is fine before the canula came out things get gradually worse over hours, not minutes.

In fact the behavior seems entirely consistent with other physical and chemical processes that are modelled using a “half-life”. This gives rise to very simple math which is widely, completely, understood and widely used.

I said exactly the opposite; the fact that we are not dead after an hour, or two or more, indicates that the insulin that was previously injected is still being effective.

This is somewhat off topic, what I said stands, I’m just trying to clarify why I said it and explain the science behind my post. One of the most important things to understand is that a site change doesn’t stop the old site delivering insulin. For a site change to be complete the insulin effect on the old site has to have dropped to zero, or close to zero; 8 hours for normal and fast acting insulin is a safe number though, in practice, it varies with the site, the specific insulin, and, apparently “day to day”.

I assume that the “day to day” variation corresponds to changes in diffusion of the insulin resulting from site changes, exercise etc. Unfortunately the product insert for Fiasp does not say what the variation in half-life is! That single figure would help us understand how much a site change can, on its own, affect BG at the time of the change; not because of changing from a “good” site to a “bad” site but because of changing to or from a “slow” or “fast” site.

1 Like

Yes, depending on how soon after the site change you do the bolus. Some people deliberately bolus after a site change to speed up the insulin diffusion in the new site; this results in excess insulin over time, but in the near term it will help with a near term (1-2 hour) meal.

Another approach is to bolus on the old site; if you know you are going to eat soon (so a pre-bolus is safe) do it before the site change. The new site then has time to “charge up” with insulin.

If you don’t eat, so have a constant basal requirement, it doesn’t matter; the old site will discharge its insulin while the new one charges up. Someone out there must have worked out a safe delay after a site change to avoid this problem. My guess is that the delay would be in the range 2-4 hours during which eating might cause problems.

This is an endo question; the math, while not as simple as diffusion, is pretty damn simple (it’s actually two diffusions in series, but there are two sites so that creates a more complex interaction, I think :slight_smile:

Someone out there must have looked at this issue…

When changing a site there are a few things that may help it keep you BG from dropping and to have the insulin start working as soon as possible. Some of these are not, for unknown reasons, not taught to new pump users.

1 - Leave the old site in for several hours to be sure the insulin infused at the site is fully absorbed, and does not leak out after removing the old site.
2 - After inserting the new site fill the cannula. The amount of insulin may vary with different infusion sets. With the T:Slim pump you can use the Load/Fill Cannula feature to ensure the the fill is not seen by CIQ as IOB.
3 - Use the Fill Cannula feature again to prime or soak the site to create a puddle of insulin at the end of the cannula. I’ve used 1u for 28 yrs with success. This creates an opportunity for the insulin to begin absorbing ASAP because as it comes out of the vial is is not effectively absorbed until the insulin molecules are separated to be absorbed.
The amount of insulin needed to fill you cannula and to prime or soak the site will vary so experimenting would be productive.

Note: If this a new site that is replacing a bad site you may need to bolus to make up for missing basal insulin because it was not absorbed in the bad site.

This should be taught to everyone who uses a pump, as I was 31 yrs ago.

Hope this can help make pumping easier.

I often use a syringe for the next meal if the site change happened less than an hour earlier. Also be sure and fill the canula even if you are using a metal infusion set like Sure-T. I do .3 for it and use Sure-T. I hope this helps.