Never trust a single lab test   (October 22, 2009)

Anytime I'm at a big diabetes meeting where there are lots of exhibitors, there's probably one or more exhibitors who offer free A1C testing. And that was true again this week, at the IDF meeting in Montreal. As usual, I took advantage of the freebie and had my A1C done, at an exhibit sponsored by a company that makes an A1C machine for use in physician's offices (Bio-Rad's "in2it").

But as happened to me last summer when I got an A1C at a different meeting, I was surprised at the result. My insulin pump has been working just fine, and I've been running a bit lower on my BG levels than a few months ago.  My A1C has been consistently below 6.0 for quite a while, so I sort of expected the A1C value would be lower than 6.0. But the result was 6.2. Yuck!

I vaguely recalled that I had wandered past another exhibit booth that offered A1C testing, and now I hunted for it, and finally I found it: MSD was promoting their diabetes pills and also offering A1C testing with Metrika's "A1CNow" device. So I had another chance to check my A1C. And this time, the result was 5.4. Wow!

I slept on the issue overnight, then decided to get repeat testing the next day. Bio-Rad's device: 5.9. Metrika's: 5.7. Hmmm. If I had not gotten the tests the day before, these look pretty good: the A1C test is generally thought to have about 0.5 points variability from test to test so that these two results only 0.2 apart are no sweat.

Now I had 4 results, and the average of the four tests was 5.8, which is about where I expected I would be. But notice what the conclusion might have been if I had only done one test the first day:

* If the only test I had done was the one that had been 6.2, I would have been misled to think my A1C was higher than previously, and I'd have to tighten up my diabetes control to get back on track.
* Or, if the only test had been the one with the 5.4 result, I would have been misled to think that I was doing just great, and my biggest concern would have been avoiding hypoglycemia.

As a result of this little episode, I find myself reiterating the exact same conclusion that I offered last summer: don't trust a single value of any lab test, anytime, anywhere, if it's different from what you expected. Instead, get it rechecked.