Tuesday, August 16, 2016

Diet and Cholesterol

Most of us with type 2 diabetes also have problems with lipid levels, both cholesterol and triglycerides. So the following blogposts about diet and lipid levels, written by a nondiabetic/prediabetic software engineer, are relevant to us.

Because the author, Dave Feldman, is a software engineer,  his blog is a bit geeky (understatement of the year), and it will probably tell some of you more about diet and cholesterol than you wanted to know. But he's done an incredible number of N=1 experiments on himself, and the results are fascinating.

He's on a low-carb diet, and in a nutshell, he's shown that at least in his case:

1. Cholesterol levels change quickly, in about 3 days, not very slowly as most people will tell you.
2. Counterintuitively, the more fat he eats, the lower his total cholesterol levels go. Also lower triglycerides, LDL cholesterol, and LDL particle number, which some people think is a better marker of cardiovascular risk than LDL cholesterol. HDL levels increase.
3. It's his diet in the 3 days preceding the test that affect the cholesterol levels. Diet on other days doesn't seem to matter.

Note that some people argue that cholesterol levels don't matter. Whether they do or don't, it's interesting to see how quickly they change with the fat content of the diet, which suggests that unless you eat the same thing every day, the lipid values you get with standard testing don't mean a lot.

After testing himself rigorously, Feldman also tested his sister. He says he's a "hyper-responder" to a low-carb ketogenic diet, meaning that when he went low-carb, his cholesterol levels skyrocketed. Although most people see cholesterol levels fall when they go low-carb, Atkins Diet author Robert Atkins had noted that in about 25% of people, cholesterol levels do go up on such a diet.

Feldman's sister, also on a low-carb ketogenic diet, is not a hyper-responder, and he wanted to see if she'd react the same way he did, so they both ate the exact same food at the exact same time of day for a few days. It turned out that although her cholesterol levels were lower, they followed the same pattern: more fat in the three days preceding the test resulted in lower cholesterol levels.

If you want the details, you can find them here:

Part I
Part II
Part III
Part IV
Part V
with more undoubtedly to come.

Feldman says he's planning to write something for the nonengineer. 

In the meantime, this suggests that if you're concerned about cholesterol levels that have changed from your last test, it might be worthwhile to see what you were eating in the three days before each of the tests, to see if that could have been a factor.

Monday, August 1, 2016

Self-Monitoring of Blood Glucose

Several years ago, some British studies claimed that there was no benefit to self-monitoring of blood glucose (BG) levels (SMBG) in people with type 2 diabetes. I blogged about the studies here, pointing out that of course the testing they studied had no benefit because they didn't also teach the patients what to do with the resulting data.

Some more paranoid patients said the studies were probably funded by the National Health Service so they wouldn't have to pay for glucose testing strips.

One study mentioned in my blogpost did find a small, but statistical, benefit to self-monitoring, and recommended that patients be taught how to use the data they got.

Now, UK researchers have published a study showing that, in fact, self-monitoring improves control when patients are guided in how to respond. The free full text of the study is available here.

The researchers used telemonitoring to guide 160 patients with hemoglobin A1c levels greater than 7.5%. The patients submitted their BG readings to a website where a physician or a nurse analyzed the data and made recommendation on appropriate lifestyle changes. Another 160 were given usual care.

Even though the participants in the treatment arm of the study submitted morning and evening BG levels only twice a week (those on insulin tested more often), their average A1c after 9 months was 7.9% and the patients in the control group had an average A1c of 8.4%. Baseline A1cs were 8.8 and 8.9% in control and treatment groups, respectively. Often just being in a study causes patients to improve their control.

An A1c of 7.9% is still too high, but the difference of 0.51% between groups is approximately the same as the reductions found with drugs like metformin. Interesting that the patients in the control group reduced their A1c by 0.5 from baseline.

Going on a low-carb diet likely would have reduced the A1cs even more, but a little improvement is better than no improvement.

Blood pressure was also lower in the group that received the self-monitoring and advice, but there were no differences in weight between the two groups.

What this study shows is that if you give patients help with interpreting their BG readings, you can improve their A1c levels in a clinically significant way. It also showed that just being in a study makes people more careful about their way of living. If you think someone is watching you,  you're more careful, even if you're not communicating with them every week.

The online method used was certainly less expensive than weekly visits to a health care person, and as we keep saying, strips are cheaper than complications.