Users of insulin pumps are at 29% lower risk of death compared with patients on insulin injections

From http://www.sciencedaily.com/releases/2014/09/140916182216.htm

A study of more than 18,000 patients with type 1 diabetes has shown that use of insulin pumps to administer insulin rather than treatment with multiple daily insulin injections results in a 29% reduction in all-cause mortality and 43% reduction in the risk of fatal cardiovascular disease (CVD, i.e. coronary heart disease or stroke).

The researchers found that patients with insulin pumps were 29% less likely to die from any cause than those with multiple daily insulin injections, and also 43% less likely to develop fatal CVD, with both results being statistically significant. Specifically looking at fatal or nonfatal coronary heart disease, pump use was associated with an 18% risk reduction (borderline statistical significance). Non-significant risk reductions with pump treatment were seen for fatal/nonfatal CVD combined and for non-CVD mortality.

Further info from http://www.medscape.com/viewarticle/831836

However, she (Lead author of the study, Soffia Gudbjörnsdottir, MD, of the University of Gothenburg, Sweden) cautioned that the patients using pen injections were different from those using pumps: the former had a slightly lower educational level, were slightly older, had a longer duration of diabetes, higher blood pressure, and more previous cardiovascular disease, although the same HbA1c. Although they used propensity-scoring to balance the 2 groups and compare them fairly, they may not have eliminated all confounding, she noted.

Asked if she thought the pump itself saves lives, she said: "I don't think it's the pump per se; it's everything that comes with the pump. In our country, pump users are invited to diabetic school when they start the pump, and it's easier to contact a doctor or nurse on weekends, so you have better care."

 

Yeah, us MDI'ers are so dumb.

But seriously, I think it is dangerous to assign causality where there may not be a connection. It is easy to toss out statistics like this but it is important to understand how it might affect people. Additionally, it is important to note that it was not an even split between the 18,000 people, with 9,000 using pumps and 9,000 on MDI. It was actually only a little over 2,000 on pumps and over 15,000 on MDI. Could that have had an effect on the percentages?

The study also did not really say WHY the conclusions they drew occurred. Also, they talk about "all" deaths. So if a person was on MDI and they got hit by a car and died, does that count towards the percentage? Another point that was actually voiced by the author of the study herself is that if those people in the study who use MDI are less educated, have had diabetes longer, have more negative health issues, it might actually be those factors, and not the method of insulin delivery, that influence the percentages.

The only reason I am posting is to assure those members who use MDI, especially the newer members, that MDI is a safe and effective way to manage their child's diabetes and to just take a rational, objective look at studies like this before making any major decisions.

 

I am wondering if there is a third variable between people choosing insulin pumps and people with diabetes who have a reduced risk of death. Maybe those that are using insulin pumps are more financially able to control other risk factors that affect CVD, particularly diet and exercise, thus reducing their overall risks of death compared to MDI?

Just because I have a psychology exam I need to study for and should be working on applying my knowledge: From the general gist of this study, I'm thinking that it is correlational, so no causal claims can be made because of the third variable potential.

 

I have met just as many can't afford to pump people as I have "could afford it" and choose to not pump. And I have met plenty of low SES familys that eat so much better than some higher SES families. There are just so many variables to be looked at.

 

Comparing 2,000 in one group to 15,000 in another group will cause zero problems, if they did their statistics in an even slightly competent manner. Reporting the wrong results because the comparison groups are of different size is a freshman-statistics mistake (maybe even a high school statistics mistake). It's not the kind of thing you will see in a published paper.

Your correlation vs. causality is always at least a possibility, but it's a very generic complaint. It could be made about almost any study at all. Do you have a specific reason for thinking that this study is reporting a correlation rather than a causation?

Yes, maybe, which is why the researcher mentioned it. However, they knew about all those issues, and tried to handle them (as confounding variables). That's the right thing to do. If they do have a problem, it is likely to be a factor similar to those other factors, but one that they didn't check for or didn't know about at all. Of course, they don't know what that would be (or they would have checked for it), and it might not exist at all.

Also, you are wrong to say that the study did not explain why mortality was higher for the MDI group. It did: heart attack (referred to as "fatal cardiovascular disease"). The level of heart attacks were much higher for the MDI group than the pump group. Since heart attack is a major cause of death, if pumps reduce heart attack risk, that would certainly cause the lower mortality seen here.

That's your opinion, and you're free to have it, of course. However, I would not present this opinion as fact. I would not dismiss this study just because you don't like the results. In general, I don't change my behavior based on one study; I like to see 3 or 4 studies, with consistent results before I'm sure they are right. However, this study looks like a good one to me, and I definitely think it is the first step on the path that leads to pumps being considered superior to MDI. It's not the end of the journey, but it is a start. And I think it's a huge mistake to "assure" people that the study is wrong, without good reason.

I'm curious what you mean by "take a rational, objective look at studies like this". What is irrational or non-objective about the conclusions of this study? They found a lower rate of heart attacks which cause a lower overall rate of death. Seems very rational and very objective to me.

This study was done in Sweden, which has some kind of single payer or national health system, so issues of affordability should be much less important there, than here in the US. I'm not saying they won't exist at all, but I'm sure they are vastly smaller than here.

Joshua Levy

 

All good points. It may be as simple as that people that use pumps in general are more proactive in dealing with diabetes and general health. The people like Christopher who are sophisticated users of MDI and also sophisticated in terms of diet and exercise are fewer and farther between than those who go to pumps. This may be even more true in this study as I believe pump use in Sweden is very heavily promoted from day one of diagnosis, and therefore those who do not pump are more "actively" not going on pumps. After 43 years of T1 for me the ability to adjust insulin with a pump is easier than on shots. It can be done on shots, but I found it much more stressful and took much more pre planning and much more eating than on a pump. The ability to reduce your basal insulin within a few hours as opposed to 12 hours is huge. I was on MDI for most of my life as a T1 and I was so sick of eating to deal with unexpected hormonal changes or a bit more exercise or a bit better injection absorption site that for me the change to the pump was huge. It may be that as a T1 as I aged my remaining insulin cells and compensating hormonal systems got worse and worse and thus MDI got more difficult. I think exercise, hormones, diet, length of time as T1 all contribute to how well MDI or a pump routine works for an individual. I frankly had such horrible hormonal impacts that MDI was always going to be a nightmare, pump use helped but it still was hard, but the pump with no other change to anything dropped my A1c by more than one full percentage and reduced the amount of food I had to eat and lows I was experiencing. I dropped over a point while reducing my lows. For me it was a huge improvement. Ali

 

I never presented my opinion as fact. You are making assumptions and misrepresenting what I said.

I am not dismissing the study. Again you are making assumptions and misrepresenting what I said.

I never said the study was wrong. Again you are making assumptions and misrepresenting what I said.

 

I think that if the various bionic/artificial pancreas efforts come to fruition, this may all be a moot point.

 

Empirical evidence that should prompt automatic insurance coverage of pumps (and dare I say, affordable access--no, wait, this is America).

 

Really? I don't think Joshua was purposefully deceiving anyone. If he is "misrepresenting" you then you very seriously mis-worded your post.

 

This is it right here. The pump users have better care.

I've seen it here in Canada as well. You want to get your child on the pump, and all of a sudden there are doctors and nurses available to give you many hours of education - not just about the mechanics of how to use a pump, but also many other aspects of D management including nutrition, exercise, and coaching about the importance of good habits and regular testing. None of this is available to those who elect to stick with MDI.

It never mattered to us because of the stellar training we got at diagnosis in Germany, but for most families this difference alone would be huge.

 

"Unmeasured confounders (such as personality, type of care, how often blood sugar is controlled, diabetes education, use of continuous glucose monitoring (CGM), adherence) could, say the authors, affect the results. However, they carried out a sensitivity analysis that showed that these confounders were unlikely to affect the findings."

The above from the article confuses me. I know that for me, it would not be possible to control my DD's morning BG without a pump, from 3-6 AM, her basal rate is increased by 300%; temp basal allows for a degree of control not possible with MDI; pumping makes it easier to correct BG periodically during the day (you can do this via injection, but I know DD would be less likely to correct if she had to inject); and the pump helps with dosing calculations and keeping track of IOB (you can do this with an external app, but I know DD would be unlikely to use an external app consistently). So for us, pumping is directly responsible for better control and better health. So, If not for the confounders the authors ruled out, what is it? The article specifically did not say pumping leads to better BG control.