Monday, February 18, 2019

Me and NPH

As everyone knows, insulin prices are now criminal. For a long time I was using Levemir, just once a day, in the morning. It worked fine, as I don't seem to need insulin overnight. If I just stopped eating, I could eliminate insulin altogether, but that doesn't seem like a likely scenario.

For more than a year, I've been on Tresiba as part of a clinical study, which meant the insulin was free. That was nice. The Tresiba worked OK, but because it lasts for 24 hours, it meant I couldn't use as high a dose as I used for Levemir. I was on 20 units of Levemir but only 12 units of Tresiba. Any more and I'd go low overnight. In fact, even on the 12 units, I was going slightly low at night for long periods, according to the Freestyle Libre continuous glucose monitor (CGM) I was trying at my own expense.

When the Tresiba trial was over, I still had two Levemir pens that hadn't quite expired, so I used them. But when I found out that refilling my 3-month prescription would cost about $1000, I figured no way and  decided to try NPH. For the last week of the Tresiba trial, we'd been put on NPH, and it seemed to work OK, but my CGM stopped working after only a day on the NPH, so I was curious to give it a longer trial.

NPH is about $24 for a 10-mL vial of 100 units/mL at Walmart. So at 20 U/day, a vial should last me about 50 days, or 48 cents a day, $43 for three months.

On my Plan D, I actually pay only $30 for a 3-month supply of Levemir, but my plan pays $1359.40, and because the infamous doughnut hole is based on total drug costs, not what the patient pays, with those prices plus another expensive drug, I'd reach the doughnut hole for sure.

Other than expense, an important factor is how well it works. Two endocrinologists warned me that NPH is  unpredictable, that the same dose might work one way on one day and another on another. But I find that diabetes is unpredictable anyway. You can eat exactly the same thing and get the same amount of exercise on two days and get different results, so it's difficult to know which of the myriad factors that affect blood glucose (BG) levels is responsible.

Dr. Richard Bernstein, low-carbohydrate diet proponent and author of The Diabetes Solution, warns about using NPH because it contains protamine, which can interfere with the reversal of the anticoagulant heparin. But an endocrinologist I asked about this said he used NPH for years before the modern insulins were available and he'd never seen this problem, and besides, people these days tend to use other anticoagulants, so I decided not to worry about it. Anyway, I'd already used it in the clinical study.

Bernstein also says one reason the cloudy insulins seem unpredictable is that people don't shake the vials enough so the insulin isn't distributed evenly.

So what did I find?

NPH is definitely peakier than Levemir. Some sources say it peaks 4 to 8 hours after you inject, others say 4 to 6 hours, and the Joslin Diabetes Center says 4 to 12 hours (at least they're consistent with the 4). So there's obviously a lot of individual variation. I find it peaks about 7 hours after I inject. I get up early, usually around 5 am, and inject then, so my NPH peaks at lunchtime. This is handy, and I try to eat my big meal of the day at noon. If I don't have lunch, I might go low.

This is, of course, not the way you're supposed to use a basal insulin. You're supposed to inject just enough to keep your BG at a good level when you don't eat and then use a bolus insulin to cover meals. I'm too lazy to do that. And as a type 2 I have a bit of a buffer because I'm still producing small amounts of insulin, just not enough to cover meals.

Some people take a second shot just before going to bed to use the NPH peak to cover the dawn phenomenon. However, although my BG goes up a little in the morning, starting when I wake up, the rise is not large.

After the Tresiba study was over, I went back to Levemir until those pens were finished. Then, wearing the CGM,  I tried NPH, 5 units in the morning and 5 in the evening. Then I tried 10 in the morning. Then I tried 15 and then 20 in the morning. The 20 didn't make me go low except for once, and that was about an hour after I injected so it was probably not a result of too much insulin. Maybe I injected too close to a blood vessel. That happened with Lantus once, and I went down to 25.

Then for two days I tried no insulin, but the CGM gave out after just one day.

Then I got a new sensor and tried a few different concentrations.

The program you can use with the CGM calculates average BG each day. Here are what I got. When there's only one dosage listed, I took it first thing in the morning.

12 units Tresiba:  98, 105, 96, 91, 92, 103, 99, 96, 92, 89 (96.1 average with overnight lows )
15 units Levemir: 98, 104 (101 av)
20 units Levemir: 107, 89, 103, 104, 82, 94, 93, 100, 106, 110 (98.8 av)
10 +10 units Levemir (10 units in moring and 10 in evening):  90, 93 (91.5 av)
0 insulin: 114
5 + 5 units NPH: (5 in morning and 5 in evening: 118, 109, 103 (110 av)
10 units NPH: 102, 106 (104 av)
15 units NPH:109, 101, 96, 101, 109 (103.2 av)
20 units NPH: 101, 100, 109, 101, 99, 95, 97, 96, 86, 103, 93 (98.1 av)

Note: One problem is that the Libre is that every sensor can read slightly differently, and I find that they read high on the first day after insertion, so I don't use that day.

Thus the Levemir given twice a day gave the lowest average BG, but that was only two days and might not be reliable. Not surprisingly, the highest average was with no insulin, but that was only one day. The sensors and then the receiver kept not working at important times.

12 units of Tresiba gave the next lowest average BGs, but there were also a lot of overnight lows.

The 20 units of Levemir given in the morning, with lots of data points, and the 20 units of NPH given in the morning, with lots of data points, gave about the same averages, 98.8 and 98.1, respectively, and the NPH readings were with a sensor that read about 10 points high.

So FOR ME, it seems as if NPH works just as well as Levemir. Whether or not I'll continue with it permanently depends on how close to the doughnut hole I seem to be getting. It is a lot more convenient to use a pen. I think NPH comes in a pen, but that's more expensive.

And my feeling is that Big Pharma won't reduce insulin prices unless people stop buying their product when it costs too much.

My experience suggests that patients with no insurance and low incomes might find that they were better off with the old-fashioned bottled NPH, with regular (R) for meals if they use a bolus insulin. They're not as convenient as the newer insulins, or the pens, but no one should die because they can't afford modern insulins. Low carbohydrate intakes require less insulin, and R actually covers low-carb meals better than the faster bolus insulins.

I plan to do more experimenting with the CGM in the future. If I have any interesting results, I'll post them.


  1. Hi its comedycook, long time. You are so helpful. This is great info.. But to be honest my head is spinning. After 20 plus yrs managing type 2 I am becoming weary. And yes, my pen insulin of Lantus and Humalog prices are thru the roof. I'm just weary and I am sure many many people are too. Just venting. Thx for all you do. Linda

  2. Linda, how nice to hear from you. I think we all get weary. We have to remember that at our age there are worse things. Hang in there!

  3. NPH worked well for millions of diabetics.