Dismal Medtronic 670g Results for Pediatrics

MiniMed 670G and Guardian3 CGM sensor from Medtronicdiabetes.com

MiniMed 670G and Guardian3 CGM sensor from Medtronicdiabetes.com

The USA Endocrine Society is singing the praises of Medtronic MiniMed's 670g insulin pump system. Although it's not yet FDA approved in the USA for children ages 7-13, a study funded by Medtronic has shown that the 670g is safe and effective for pediatric patients with type 1 diabetes.

As you read below and watch the video presenting this data, you'll see why I'm not convinced this high-tech "smart pump" system is making ANY notable improvement in the health of the people using it. 

Medtronic's 670g is touted as "THE WORLD’S FIRST HYBRID CLOSED LOOP SYSTEM."

It was approved in the USA for adults and adolescents 14 years and older in September 2016, and released for that age group in April 2017. 

The 670g works directly with the new Medtronic Guardian3 CGM system to detect and automatically react to your current blood sugar levels. Using the 670g automatic mode, the user sets a standard blood sugar range and the pump will automatically suspend insulin delivery 30 minutes before the user reaches the low threshold of that blood sugar range, and the pump will automatically bolus more insulin when you approach the high end of the preset range. 

I originally saw this research summarized in print form here and I needed to do some digging to find the original presentation (video below) because the data given in print wasn't adding up. 

While the data DOES show an improvement in overall A1c and blood glucose variability within children ages 7-13 years old using the new Medtronic 670g system in automatic mode, it's a measly, little, tiny, minuscule improvement. 

Is this even noteworthy? 

Listen to the presentation below [starting at 11:10] describing this study using the Medtronic 670g for the pediatric group (ages 7-13),  or skip to a linked synopsis with highlighted quotes below from the article: 

The data shows...

"In a single-arm study, Wood and colleagues analyzed data from 105 children aged 7 to 13 years with type 1 diabetes.... Following an in-clinic evaluation, children used the MiniMed system in open-loop, manual mode for 2 weeks, followed by closed-loop, automatic mode for 3 months.

From baseline to 3 months, HbA1c improved from a mean of 7.9% to 7.5%, as did overall variability of sensor glucose values. Total daily dose of insulin increased slightly, from a mean of 0.8 units per day to 0.9 units per day.

[ diaVerge interjects: That improvement from 7.9% to 7.5% A1c is an average improvement of only 14 mg/dL. That's from 204 mg/dL average to 190 mg/dL average.] 

Wood noted the researchers were most excited by the marked improvement in time spent in the target blood glucose range of 71 mg/dL to 180 mg/dL, which increased from 56% to 65% in the pediatric group over 3 months of use."  

[diaVerge interjects: Since there is roughly 720 hours in a month, this would be an increase from 403.2 hours in range to 468 hours in range per month. That looks like a lot, but when you consider the huge target blood sugar range of 70-180, an added 64.8 hours per month within that range AND still maintaining a blood sugar average of 190 mg/dL (which is outside the pump's target range) is atrocious. 


And the most shocking detail is, not the huge target range of 70-180 mg/dL, but the standard deviation 'improvement' shown in the chart above in yellow (slide taken from the video presentation above, with yellow highlights added by diaVerge).

The standard deviation is the range that your blood sugar fluctuates on either side of your average. 

Across the age ranges, you see that the standard deviation of blood sugar levels is still very high, and with a very marginal improvement using the Medtronic 670G smart pump system.

As you see highlighted in yellow, the pediatric group standard deviation improved from 67 to 63 mg/dL, Adolescents from 63 to 59, and Adults from 54 to 51 after 3 months using this pump system on auto-mode. 

This automated pump system IS an improvement and, no doubt, this technology is a godsend for many people. A pump system that can take much of the guesswork out of insulin delivery and do so automatically - with no user input - is amazing. 

But... we deserve better. 

We deserve lower blood sugar level thresholds AND a lower standards of deviation.

This system is NOT going to ensure the long-term healthy outcome that we are all working towards, and especially not in children who have a long and arduous future with type 1 diabetes ahead of them.

It makes me angry that so much money is put into developing 'tools' to make diabetes more convenient. This is a patch instead of a cure. 


It's no surprise that many people following the low carb method of diabetes management are opting to NOT use this pump system because we can manage our blood sugar levels better on our own.

I've spoken with a few low-carbers who DO use this Medtronic pump system in manual mode only (not on auto-mode) which makes it just like other manual pumps with integrated CGM data.

Most people with type 1 diabetes eating low carb, following Dr. Bernstein's plan for diabetes management, can maintain standard deviations in the teens and low 20s (compared to the data above, which shows standard deviation from 51-63) 

But don't take my word for it. Here's proof from people with type 1 diabetes, who follow low carb ala Dr. Bernstein and maintaining NORMAL (non-diabetic) blood sugar levels: 


(A huge thanks to all the people who provided me with CGM data to use for this article. More graphs yet to be added. ) 


We do NOT NEED "smart pump" that maintains HIGH blood sugar levels and in fact, provides only marginal improvement. Low carb is more effective at lowering overall A1c levels and standard deviation than the Medtronic 670g.