Our last 36 hours were very frustrating due to a site going bad on day three—except that we did not recognize it as such.
The primary symptom was that basal needs suddenly increased, over a few hours, from -15% (where they had been for a week), to +20% (which is typical of upcoming sickness).
This basal change, of course, came with quite a bit of trouble, because, for us, adjusting to basal change takes typically 8-12 hours during which our control is often poor. In this case, it was very poor.
After the first twelve hours, we had adjusted to the new basal level, when we started getting violent peaks and valleys without a clear cause. Unfortunately, for a teenager in puberty, this phenomenon is not unknown–it is actually rather frequent (at least for a peak). After a few hours of this regimen, we started suspecting a site problem. We were getting close to the 72 hour period for this pump site (Omnipod), so we were due for a change, but stupidly we kept it on for a few hours longer (we often keep a pod going for another 4-8 hours for schedule convenience).
This delay led us into a very long 300 peak that took large amounts of insulin, much more than the correction factors would indicate (about 14 units over 6 hours). We changed the pod when the peak was finally going down.
The interesting part of it is that, after the new pod was in place, taking care of the insulin tail for this peak was an absolute bear. Our insulin tail is normally 5.5 hours (although, naturally, when you have injected 15 or 20 units for a peak we see a longer tail). But, in this case, we were still having to deal with delayed insulin absorption 10 hours after the insulin had been injected.
Not being sure of where we really were in basal after all that trouble, we had another few hours blundering along for a new basal level after having changed our pod, which resulted in a second (but easier) peak of long duration. It is only when injecting for that peak that we were able to prove to our satisfaction that the new pod was getting a timely response to an injection, thereby proving that the old pod had been the problem all along.
But this cost us 28-30 hours of terrible control with the old pod, plus another 8 hours with the new pod, and a whole night of staying up (at one time, we were hand testing every 5 minutes because we were chomping large quantities of fast carbs every 10 minutes and still going down dangerously low). After all was said and done, we are now back to the basal level that we started with.
I am looking back and wondering how we could have spotted this earlier and avoided this pattern altogether. The problems:
- the corrections still worked, just not very well
- the basal still worked, just not very well
- sickness creates the same elevated basal pattern, as well as the same insulin insensitivity in corrections
- until 8 hours into the new pod, there was no absolute proof that the old site was progressively failing
What is the rule you use to diagnose a site going bad? Until now, I have used a somewhat similar rule to @Jen 's (two failed corrections on a row and one pen correction working), but this did not work out well here.