Sunday, January 29, 2012

Sensor Data For 100 Days with more than 12 hours' of readings

I finally have a hundred days with more than 12 hours' of reading on the same receiver, and I recently noticed that on the dexcom daily statistics chart, I can click on the category and it sorts my days. I started doing all the computations below when I'd had a day under 200 and started wondering- how often do I have days where my high is 200 or less? 18% of the time, as it turns out- between once and twice per week.

Including all days since October 15 with more than 144 readings; excludes November 10 (no readings), October 15 (10 readings), today (77 readings as of the upload) November 27 (96 readings), December 3 (99 readings), November 11, 131 readings, and December 12 (139 readings).

Of the remaining days,
one has more than 288 readings (daylight savings day),
three have at least 288 (missing no data),
fifty four have at least 276 (missing less than one hour of data),
sixty seven have at least 264 readings (missing up to two hours of data),
eighty have at least 252 readings (missing up to three hours of data)
eighty five have more than 240 readings (missing less than four hours of data)
eighty nine have more than 228 readings (missing less than five hours of data)
ninety have more than 216 readings (missing less than six hours of data)
ninety two have at least 204 readings (missing up to seven hours of data)
ninety four have more than 192 readings (missing up to eight hours of data)
ninety five have more than 180 readings (missing up to nine hours of data)
ninety eight have more than 168 readings (missing up to nine hours of data)
ninety nine have more than 156 readings (missing up to ten hours of data)
one hundred have at least 152 readings (missing up to 10.20 hours of data)
seven days have less than 144 readings (missing at least 12 hours of data).

The lowest max value was 159; the highest was HIGH. The median daily high was 237.5.
1 day was entirely below 160.
16 days were at or below 200; 15 days had highs between 160 and 200.
53 days were at or below 240; 37 days had highs between 201 and 240.
79 days were at or below 280; 26 days had highs between 241 and 280.
93 days were at or below 320; 14 days had highs between 281 and 320.
96 days were at or below 360; 3 days had highs between 321 and 360.
99 days were below 400; 3 days had highs between 360 and 400.
1 day had a reading of HIGH.

Both the lowest high and the highest high were checked against my meter. My accu chek reading for the "159" was 164; the accu chek on the "HIGH" was 302. 159 is reasonably close to 164 and I wouldn't be surprised either way if I had or hadn't been above 160 for that day. I checked the blood sugar after a reading of HIGH had been on my Dexcom for an hour, so it's possible that it was coming down by then and that I was really above 400 for some part of that hour, but my guess is that I wasn't really.
The runner up day with the highest sensor reading had a 389 about half an hour before I woke up, and then went to ???. Half an hour later, my bg on my meter was 250. I strongly suspect that the 389 was wrong and that I was not above 300, let alone 360. Could be wrong. The other two days with sensor readings above 360 may have been accurate; one I checked and my meter said 356, the other I took a correction for without checking my blood sugar, and I didn't go hypo.

Min daily value ranged from LOW to 110, median daily low was 72.
84 days did not have a LOW; 16 days had low values of LOW.
68 days stayed at 50 or higher; 16 days had low values of 40-49
56 days stayed at 60 or higher; 12 days had low values of 50-59
29 days stayed at 70 or higher; 27 days had low values of 60-69
13 days stayed at 80 or higher; 16 days had low values of 70-79
7 days stayed at 90 or higher; 6 days had low values of 80-89
2 days stayed at 100 or higher; 5 days had low values of 90-99
1 day had a low value of 105, and 1 day had a low value of 110.

Daily average ranged from 99 to 201, median 140.
Inter quartile range varied from 28 to 144, median 60.5. The interquartile range of the interquartile range was 48 to 75.
The standard deviation ranged from 20 to 110, median 49.5.

Saturday, January 28, 2012

New Supplies

I ordered new supplies last week. I received 3 boxes of 90 syringes, 2 boxes each with 1 vial of Novolin R, 1 box with 5 Lantus Solostar pens, and 8 boxes containing one vial of fifty Accu Chek Aviva Plus Test Strips.

The Accu Chek Aviva Plus test strips are new to me, although I heard about them a few weeks ago- the person I heard about them from is in the UK where apparently they've been out for a while. They use a different blood sugar measuring technique, one that doesn't cause falsely elevated readings in people who have high blood levels of non-glucose sugars. I'm a bit concerned about whether this means that the strips will have a different level of accuracy overall as compared to the old strips. I may do a comparison run between strips at some point but not now. Here's a picture of the boxes:

It's always interesting to me to see what insurance asks me to pay, and what they pay, given that it's a new year and they're always changing these things. This year I seem to be paying a flat fifty dollars per prescription. So for the syringes, I paid $50 and insurance paid $10.41; for the Novolin R I paid $50 and insurance paid $75.41; for the Lantus I paid $50 and insurance paid $146.39; for the test strips I paid $50 and insurance paid $342.02.
I'll probably be getting four shipments of syringes this year (maybe five, actually- my prescription isn't for enough syringes), four shipments of Novolin R, three shipments of Lantus (I am supposedly getting three month supplies, but they send me a box of five pens and each pen lasts a month), four of test strips. I have a stockpile of pen needles that will last me a year at the current rate of one per day, and I have enough lancets to last at least another year at my current rate of about one per day. I will probably be getting another glucagon this year. I pay for my testosterone without insurance. So that means that if I use the same meds this year as last year, my total cost, not including Dexcom or testosterone, should come to $800. That's affordable.

Without insurance, the costs for these things would be somewhat different anyhow, but at these rates:
the syringes cost $20.17/box and I'd use at least a box per month ~ $245 for the year
the Novolin R cost $62.70/vial and I'd use roughly eight per year ~ $501 for the year
the Lantus Solostars cost $39.27/pen, and I'd use one per month ~ $472 for the year
the test strips cost $49 per vial, and I'd use at least 24 per year ~ $1176 for the year
one glucagon kit is about 90
So without insurance, this lot would cost me ~ $2484, still doable, but not easily. Although of course if I really had no prescription insurance, I would be able to get everything except the test strips for free, so my costs would actually be just the $1176.

By the way, I've finally written up the November facts for this year and put them with last year's on a page you can view above the posts.

Thursday, January 26, 2012

The False Premise Underlying Glycomark

The GlycoMark company is a joint venture which, according to its website, is "dedicated to the discovery and commercialization of novel biomarkers for diabetes."

So far it has marketed exactly one such biomarker, often called the GlycoMark test. It has been on the market for roughly twenty years but I've only known a couple of people who tell me their endos order the test regularly; mine never has. The test sort of interested me when I first heard about; you can test just the time spent high? Well, sort of. The wikipedia article on the GlycoMark test* (and the wikipedia article on the A1c test*) read like an ad for GlycoMark. I will summarize what they say.

According to these articles, there is a substance in the body that is excreted when the body excretes sugar, which is at 180 mg/dl. Therefore, people who spend much time with blood sugar have lower levels of this substance (1,5 AG). Therefore, the Glycomark test can tell you if a person's A1c of, say, 6% is because the person had near constant blood sugars between 110 and 170, or if the person's 6% represents blood sugars bouncing between 60 and 320. A Glycomark score in the normal range means the person has blood sugars below 180 almost all of the time, and progressively lower glycomark scores mean more and more time above below 180.

The Glycomark website hints at the problem with this interpretation of the GlycoMark. First of all, if the glycomark score declines with time spent over the renal threshhold, great. But to say that the renal threshhold is "usually >180 mg/dL" is disingenous. Yes, the renal threshhold is often above 180 (although a renal threshhold of 160, say, would not be unusual). But the renal threshhold can also be above 280. And there's TREMENDOUS variation in renal threshhold, between individuals, and even in the same individual at different times.
A secondary and less problematic issue is that the glycomark depends on something which has a wide normal range to begin with. Two nondiabetics could have significantly different glycomark scores.

And while knowing whether or not your blood sugar has been above your renal threshhold MIGHT be important (there's some evidence that kidney damage happens at that threshhold), you can find that out by testing your urine for sugar. I bought 100 urine sugar test strips last year for $15. One glycomark blood test costs, online, $53. Rough choice? I don't think so.

*As accessed today, Jan 26, 2012.

Monday, January 23, 2012

When Do Complications Matter?

One thing I can't help noticing in reading about diabetes complications is this: when researchers or writers want to use real data to make it sound like most diabetics don't have x complication, they talk about the most severe forms of it. For instance, instead of talking about diabetics with retinopathy, they talk about diabetics who are blind, have proliferative retinopathy, or whose retinopathy is rated severe. On the other hand, when researchers want to talk about how common some complication is, they include every form of it, for instance, mild retinopathy found on a fundus camera, or even "pre-retinopathy".

This confuses a lot of people. One prominent member of the diabetes online community has repeatedly used two studies to claim that diabetes no longer threatens lives, in a totally erroneous way. One study looked at the frequency of end stage renal failure in diabetics diagnosed at various ages, over roughly twenty years; the other looked at risk of death in diabetics in the first twenty years past diagnosis based on the presence of any kidney disease.
The observed deaths for those with no kidney disease whatsoever in the first twenty years past diagnosis was pretty much the same as for nondiabetics. The observed incidence of kidney failure in the first twenty years past diagnosis for those diagnosed under the age of 5 was nonexistant.
So this guy says, tada, diabetics without kidney disease don't die early, our kids don't get kidney disease, therefore they don't die early.

Problem? The first study defined diabetic kidney fairly broadly (they could have probably been even more broad) as any two positive microalbumin urine tests over the study period. That included just over 70% of the group. A very small portion actually had kidney failure.
The second study didn't look at kidney disease, just kidney failure.
The other problem, of course, was that it only looks at a twenty year period. If I live 25 years past diagnosis, I'll be dying young. Diabetes is definitely capable of killing a person 30, 40, 50 years past diagnosis.

Anyways, another online person recently ranted about somebody who said that retinopathy was common in type 1 diabetics, about how that person was an ignorant so and so, that these days it's not. Well... retinopathy is common in type 1 diabetics. Even in teens. Averaging a handful of European, American, and an Australian study on what portion of teenagers living with type 1 diabetes have retinopathy as shown by a fundus camera, the rate is roughly 1 in 7 boys with type 1 diabetes and 1 in 11 girls. And that's been pretty stable for over twenty years. Looking specifically at studies of teenagers who'd had diabetes for more than five years, the rates jump up to more than one third of teens with type 1 diabetes. Looking at adults, the numbers are considerably worse; in most studies, by fifteen years, the majority of participants have retinopathy.
This was even true in a study of Swiss adults diagnosed as young adults/older teens, diagnosed less than 20 years ago, tested for retinopathy 12-15 years past diagnosis. Their average A1c was 7.0% and the portion with any retinopathy was 61%.

The overwhelming evidence is that an A1c in the recommended range is not enough to prevent some complications, even though it makes them slighly less likely to occur and far less likely to progress to the point where it, say, makes you blind.

So the question in my mind, as I read these studies, is this: if I develop mild retinopathy, does it matter? As long as it doesn't impair our vision, should we care? The majority of type 1 diabetics, even with the best control they are likely to be able to achieve, develop some degree of retinopathy. Detecting retinopathy is of some importance because some types can be treated to prevent vision loss. But is it important to know about retinopathy if what you've got is very unlikely to progress? Why bother testing for the most minor degrees of retinopathy? Why look at "pre-retinopathy"?

Similarly, there's a nerve conduction test that's sensitive for neuropathy. It hurts. So the question is, what's the point of that? After enough years with diabetes, most of us will show up as impaired on a nerve conduction test. If it doesn't bother us, why test it? And if we are symptomatic, well, 80%-90% of us have some neuropathy after twenty years, so why not go ahead and pretend we've alredy done the darn test? Isn't what matters whether I can feel the things I need to feel and that my nerves don't hurt?

There's a part of me that thinks it's interesting that diabetes affects bodies in such difficult to see ways. A part of me that thinks it's neat and worthwhile to know all the ways in which my body is likely to be different from the bodies around me. There's a part of me that hopes that this knowledge could explain some cool things. But I have a large suspicion that the researchers will not use the knowledge that early diabetic changes can be seen in kidney biopsies or that various tests show diabetes in eyes in order to praise the variety of God's creation. I really don't think that's what they have in mind at all.
I figure that what they'll use it for is
to show that the way we
take care of our bodies
is not good enough,
that we need
more medical services
that only they can provide;
more self control
more shame
that we are
not as good.
Even if
we thought
we were doing
just fine,
they say it's
not good enough.

So I ask you;
if my eyes see,
are they good enough?
if my kidneys pee,
are they good enough?
if my heart beats,
is it good enough?

Can I get a second opinion!

Is the goal of diabetes treatment,
to make my body just like your?
Is the goal of diabetes treatment,
to make my blood sugar normal?
Or is the goal of diabetes treatment,
to keep me living long and feeling good?

Can I get a second opinion!

FYI, I like my doctors fine and have no diagnosed complications, nor any worth diagnosing.

Sunday, January 15, 2012


I've been astonished lately by how many insulins are sold by the same company under different names in different companies. Some examples:

Novo Nordisk sells insulin aspart as NovoLog in the United States and as NovoRapid in Canada, the UK, Israel, Australia, and I'm not sure where all else. It seems, going by its website to sell it simply as "analogue de insulina" in some South American countries but I'm not sure I'm reading that right.

Novo Nordisk gets even sillier the way it sells regular. It sells it as Novolin R here in the US and in some other countries. It sells is as Novolin ge Toronto in Canada (okay, I get it- Canadian pride). It sells it as Actrapid in Australia and some other companies. It sells it as Novolin S (just to confuse things further-the S is used to stand for Suspension and is used by some other insulin companies) in other countries. That's four names for one insulin!
I am not certain if Actrapid HM is the same as Actrapid or if the HM is productive. The standard Actrapid- like most Novo Nordisk insulins- is made by baker's yeast. But I'm not sure if the Actrapid HM is instead an animal insulin modified to be human insulin at the molecular level- some other insulin companies do this with pig and cow insulin, so maybe. Does anybody here know what HM means after the name of an insulin?

Porcine insulin in zinc suspension is sold by Merck-Plough-Shering as vetsulin in the United States (or it was, until a few months ago) and as caninsulin in Canada, the UK, Germany, Italy, Spain and Switzerland.

Wednesday, January 11, 2012

Yesterday I tried to make photocopies of the zine, but my parents' photocopier kept jamming the pages (especially frustrating you're doing two sided copies). So I decided to go find a copy shop. My mother suggested a kinko's about a mile away that she wasn't sure was still in business. Guess what? It's not there. But the library was only another block away and it had a big BOOK SALE sign so I went in. Found an old (1962) book The Genetic Code by Isaac Asimov (Isaac Asimov is the answer to: Who has written books in all ten Dewey Decimal categories?) Picked it up and looked through it, was amused to see page 92, which says:

Suppose you had a peptide chain of thirty amino acids, as in insulin.

I initially interpreted that to mean that Isaac Asimov thought that insulin is a peptide chain of thirty amino acids, but writing this, I think maybe he means the factual statement that insulin contains a peptide chain of thirty amino acids (insulin contains two peptide chains, and the B chain in humans has thirty amino acids. The A chain has twenty one, so insulin has a total of fifty one amino acids).

Wednesday, January 04, 2012

Numbers Outa My Ass

My blood sugar has been kind of rotten these last few days and my Dexcom is not helping. I changed sensors earlier this week and the new one is inaccurate so far. For instance, it said I was 188 rising before supper and I checked on the accu chek aviva and got a reading of 108. A re-check gave me a reading of 105. After supper, it showed me going up to about 240 and then drifting downwards. I took my Lantus. It was one of those shots- you know the ones, where you stick in the needle and meet resistance and push through and it doesn't hurt but then you meet another layer of resistance and push through that and then it hurts like hell? Yeah, it was that kind. I did the shot there anyways and when I pulled out the needle my butt bled a lot. So out of curiousity, I took my blood sugar with all that blood. I was wondering- would having insulin possibly in there with the blood or the fact that it was coming from, you know, my butt, matter to the accuracy? The meter read 303. Dexcom was saying 228. I rechecked on a finger and got a 297. Yeah, my butt beats my Dexcom.

Tuesday, January 03, 2012

You know my retroactive ISF system? So I decided to put in x as the BCR and y as the BPR and z as the ISF. And write a bunch of equations, like I ate 96 grams carb and 7 grams protein and injected 11 units and blood sugar fell 121, so (96x+7y +121)11 =z. And write equations like this for a bunch of mornings (as anybody who knows linear algebra knows, to solve systems of equations like this you need three equations), and then solve. Problem? I did this for four first-meal-of-the-morning meals in the past week and solving the system gives me answers that are totally absurd. Like protein lowers my blood sugar. Or maybe that's not as absurd as it seems to me, since I'm looking at it over a four hour period- maybe the protein slows digestion and doesn't raise blood sugar 'til later? Or maybe the low amounts of protein I'm eating mean that protein doesn't raise my blood sugar at all.

I've got five subscribers for my zine now but none who are readers of this blog. Maybe I downplayed it too much. Seriously, I'm proud of it. Also willing to offer to just mail it to you without you sending me a self addressed envelope as long as I don't get inundated with requests. One of the requests I've received so far is from Canada.

Sunday, January 01, 2012

New This Year- Zine

I am making a zine made mostly of things already on this blog with some new material and illustrations. I expect the first issue to be completed this week. I will be happy to send you the first issue -free!- if you mail me a self addressed stamped envelope. To get my address, send an email to diffabled at hotmail dot com. The zine will be pages folded in half, not thirds, so you may want a slightly larger than normal envelope. Or I can fold your issue with an extra fold. It will be the weight of 5 sheets of paper plus a rubber band.
If you feel that you've really contributed to this blog (you know who you are) and should get a copy without having to send me a self addressed envelope, send me an email. I'll probably agree with you.

For subsequent editions, the price will be either a contribution in the form of your submission or five dollars for one issue or twelve dollars for three issues.

Submissions need to be on a 8.5"x5.5" paper- cut or fold in half a standard sheet- and they need to be legible, factual, and interesting. Also related to diabetes. Suggested topics include: little known factoids, interesting stories, puzzles, or comparisons of products. Answers to questions posed in your article, sources for facts, and permissions should be included on a separate piece of paper. Please include permission from your friends and family members if you write about them.