Lipohypertrophy is one of those big words for bad side effects of diabetes that no one ever warns you about.
Basically, it’s a condition common in those of us who repeatedly inject ourselves – an accumulation of fat and scar tissue under the skin causing lumps that are not only unattractive, but interfere with insulin absorption, making it even harder for us to manage our condition. Ugh!
Honestly, we never thought we’d ever have anything good to say about lipohypertrophy… until hearing about a recent study at Stanford University with surprising results:
Researchers found that (get this!) CGM sensors placed on spots with lipohypertrophy showed equal or superior accuracy across all glucose ranges to the sensors placed on more “virgin” skin.
That was my reaction, too, so I took a little time to ask some questions of Daniel de Salvo, a post-doctoral fellow in pediatric endocrinology at Stanford and first author on the abstract presented at the recent ATTD conference. (The full manuscript on this study report is in the works, and the principal investigator and senior author is Dr. Bruce Buckingham.)
Note that Daniel is not only a researcher, but a longtime type 1 himself, who can found on Twitter as @MDwithT1D.
DM) Is this the first study of its kind?
DDS) Yes, this is the first study evaluating the effect of lipohypertrophy on CGM sensor performance.
When and where did the study take place?
The study took place at Stanford University and University of Colorado, Denver. Enrollment began in November 2013 and the final study visit took place in December 2014. The study lasted for four weeks with weekly study visits taking place at the Clinical & Translational Research Unit (CTRU), but otherwise patients were going about their daily lives at home, school, and work between visits.
Your abstract states that “precision analysis was performed on matched sensor glucose pairs in lipohypertrophied and normal tissue.” Can you explain this analysis method in laymen's terms?
Sure! Two sensors were placed on each subject: one in normal tissue and one in an area of lipohypertrophy. CGM readings were available every 5 minutes. Precision analysis was performed by comparing the sensor reading in normal tissue to the sensor reading in lipohypertrophy at each time point (i.e. every 5 minutes). A low coefficient of variation (PCV) or precision absolute relative difference (PARD) indicates a smaller difference between the two groups (i.e. improved precision). The PCV and PARD compared well to published reports of precision analysis of two sensors in regular tissue. This indicates that the sensor in lipohypertrophy was not far off from sensors in normal tissue.
What was the hypothesis? That lipohypertrophy would negatively affect both pump sites and CGM sites?
We know from previous studies that lipohypertrophy can interfere with insulin absorption, adversely affecting blood glucose. We hypothesized that lipohypertrophy would have an adverse effect on CGM sensor performance as well.
Was the research team surprised by the results?
Yes, we were quite surprised by the results! Sensors in lipohypertrophy showed equal or superior accuracy across all glucose ranges. That was definitely a pleasant surprise.
What impact might these results have on the bigger picture of diabetes care?
The evidence is convincing that sensor-augmented therapy can improve diabetes care, and closed-loop systems (requiring pump and CGM) will hopefully become commercially available in the not-too-distant future. Finding areas for pump and sensor can be a challenge, especially for those who have had diabetes for a long time. As clinicians, we are always telling patients to avoid lipohypertrophy, which can be very difficult for some. Now, we can say: ‘Avoid areas of lipohypertrophy for insulin delivery, but you can use that favorite site for sensors.’
How will these study results be disseminated and used? i.e. will CGM manufacturers include this info in CDE training? Or might they pursue further studies?
CGM manufacturers including Dexcom and Medtronic were very interested in these results. It remains to be seen how CGM manufacturers will use this info. It’s unlikely they will use this info in CDE training until long-term data is available.
What would you personally want to say to the type 1 diabetes community about what you learned in this study?
The message is simple: “You should avoid areas of lipohypertrophy for insulin delivery, but it may be OK to use them for CGM sensor sites." Those of us with T1D have limited real estate for injections/pumps/sensors, so this opens some space up! Further work is needed to assess the potential risks of using sensors at lipo sites over long periods of time.
Thanks Daniel & the Stanford team! It sure is nice to get some good news about the bad stuff once in a while!