this is a complicated question that I am not sure I can answer with complete accuracy but I will try, frequent blood sugar checks and ketone testing will help the problem from arising I was concerned about this as well when switching my son over 2 years ago to a pump, without working insulin or partial delivery bs levels can rise within an hour or two but would take several hours to develop into dka the child usually will have symptoms to tell you there is something going on, belly ache, sick feeling, frequent urination, thirst my son has only had one even close to DKA in 2 years pumping he went to bed with a normal bs got up around 12 to pee but didn't check bs like we normally would he had had a lot to drink around 130 he vomited his bs was high( over 500) and he had mild to mod ketones his canuala had totally kinked he came down gradually with an injection the doc on call really underestimated how much to give him we discussed it and basically have him do his correction for a bs of 500, lots of fluids as much as we could with vomiting but we were able to keep him at home so during the course of a normal day you are going to catch this before it escalates if he is unexplained high we try one correction if no result in two hours we usually change he pump out or at least check for ketones and maybe try one more correction. during the night is trickier of course my son hates night testing so I watch for signs, up to pee that things might be awry oh and when very high insulin resistance increases so that is why they need a bit more .....
We had a problem 3 days ago, when we started getting higher values > 300 mg/dl, just one day post cannula change.It started when my five year old son's prelunch values were below 200, but in the evening when he came after playing his value was 516. We were shocked, we gave correcting dose assuming it to be high before exercise and increasing thereafter(paradoxically which happens after exercise). We started checking after every two to three hours and all were above 300.So we were confused what dose to give ? Increase dose by 1.5 to 2.0 times assuming DKA. Urine ketostix strips did gave some indication of ketones. Finally we shifted to pen rather than continuing on pump.When his values came under control the next day , we changed cannula to another site. What went wrong is still a dilemma for us. Fortunately no symptoms suggestive of DKA developed.
We had a similar incident the other morning, when Eric apparently pulled his site out after his shower and went all night with no insulin.
Rule of thumb from my endo is, assume site failure if you get an unexpected high like that and never correct it using the pump — use a syringe. If you had low ketones, don't worry too much about DKA after you've switched to injecting insulin — as long as he has insulin in his system, ketogenesis will be suppressed. And, my endo says it takes a day or two of continuous high BG/high ketones to turn into DKA, so this is why, if you get a very high BG, you correct it with a syringe and change the pump site immediately.
When correcting for a high BG with ketones, the protocol I've been given is this:
1) if he has trace or low ketones, correct the high by injecting the amount of insulin his pump says he needs for it. I use the pump for calculating the correction, then bolus the insulin into the sink and give him the amount, or as close to it as I can get, via syringe. That way, the bolus is in the pump's computer but I don't have to worry about whether it got through a probably kinked infusion set.
2) if he has moderate ketones, give the correction dose + 10% of his average total daily insulin intake. Eric is six, so he gets on average 14-15 U/day of insulin, so I would generally add 1.5U to whatever the correction dose is.
3) if he has high ketones, I'm supposed to call the clinic, but that's mostly so they can review his symptoms and be sure they aren't dealing with something bigger and more nasty than just an accidentally removed site. What they generally tell me is, give the correction dose + 15% of his average daily total intake — so instead of 1.5U, I'd add 2U to the correction dose. This is what happened yesterday — he was at 460, his pump called for a 3.1U correction, the clinic told me to add 2 U to that, and I wound up giving him 5 U total by syringe.
So, talk to your clinic and ask if they agree with this protocol (different clinics use different techniques).
Another thing to take into account is, how long has the insulin been in his pump and exposed to heat? Eric's BG came down after his wicked high yesterday but then went back up into the 300s, and I couldn't figure out why until I realized that the reservoir he had in his pump had been filled on the 25th, and had therefore been exposed to the summer heat for five days. I can get away with using the same reservoir that long in winter, but not in summer. I replaced the reservoir with fresh insulin and by dinner time, he was 113... and this morning, he was low (argh). So after you change the pump site, do a quick check to make sure the insulin in his pump isn't more than 3 days out of the vial, and that also may help.
Sometimes we too think hot weather can cause some decrease in potency of insulin but our endo doesn't think so.He says insulin can stay at room temperature for about a month without much loss of potency.What temperatures was the insulin exposed too? and what are the recommended ones?
I have to wonder about this. There's room temperature, and then there's "room temperature". Let's define our terms precisely before reaching too many conclusions. For example, our house has inadequate air conditioning and it's not unusual for my office to stick at 85° for several hours at a time. I really wonder what that would do to unrefrigerated insulin over time.
That's true IF room temperature stays below about 75 degrees. I frequently leave the insulin out of the fridge in fall/winter when we keep our house temperature at 68 or below, but as soon as it rises above 70 I keep it refrigerated. Most of the manufacturers say very clearly that long-term exposure (a couple of days) to anything above 80 degrees will spoil the insulin. My endo also says it's more stable than the manufacturer says it is, but that has NOT been my experience. My experience is, when the outside temp is above 75, if I have the same insulin in my son's pump for more than 3 days, it loses potency. The test? I put fresh insulin in the reservoir and voila, stubborn highs go away. So with all due respect to my endo (who I adore, don't get me wrong), I'll go with my experience over what his lab findings say any time.
I also keep in mind the fact that the pump rests right next to my son's body. Now, that doesn't mean that the insulin gets to body temperature — it is somewhat insulated by the plastic and the cloth between him and it — but it does mean that there's another source of heat potentially affecting the stability of the insulin. So on a hot day, it's a bit of a double whammy for the insulin. I am not at all sure that those who assure me that insulin is so very stable take that extra source of heat into account.
sounds like a very scary night, I agree with Elizabeth, always suspect the site with unexplained highs and change out when there is any doubt, not worth the worry, especially overnight, I hope you have a better stretch of luck now!
amy
I keep my insulin in the fridge all the time…and I live in Canada! 
Is there any negative to injecting cold insulin?
Aside from the injection feeling cold, I doubt it. I’ve heard that filling a pump cartridge with cold insulin results in more air bubbles and so for that it’s probably better to use room temperature insulin. But insulin can stay out of the fridge for up to a month after it’s opened. And, considering that those of us with pumps sleep under the covers with the pumps and tubing right next to our bodies, I wouldn’t worry about temperatures unless it’s really hot out or the pump or tubing is exposed to a period of direct sunlight.
As for the DKA question, the longest I’ve gone without insulin is five hours, and I was feeling extremely sick and had very high blood sugar and ketones. I fell asleep once without my pump and woke up three hours later to find that my blood sugar had gone from about 5 mmol/L to 20+ mmol/L and I had ketones, but was just beginning to feel sick. The same has been true when I’ve had infusion sets not work properly, three hours seems to be about the timeframe that it takes for my blood sugar and ketones to get quite high, but I think it would take a few hours more to actually get to DKA. Given my experience with five hours, I have no doubt that if I were to spend an entire eight-hour night without insulin I’d wake up in DKA.
Re: storing insulin in the fridge . . . I have always done that, since day 1. Never a problem.