Background and Aims

Higher bolus frequency is expected to correlate with better glycemic control; however, there is little real-world data quantifying this trend. This study retrospectively assessed glycemic outcomes stratified by bolus frequency for a large cohort of adults with T1D using the Omnipod® Insulin Management System (Insulet Corp., Acton, MA) with an integrated BG meter (Abbott Diabetes Care Inc., Alameda, CA) and data management system (Glooko, Mountain View, CA).

Methods

Insulin pump data uploaded to the data management system from February-August 2019 were matched via device serial number to a second database of self-reported demographic data and de-identified. Data from ≥3 mo of system use per user were analyzed. Glucose Management Indicator (GMI) and percentage of readings <54 and 70-180mg/dL were calculated based on ≥14 days of BG meter readings for users grouped by average bolus frequency (<3, 3-4.99, 5-7.99, or ≥8/day).

Results

In 4,483 adults aged ≥18y with T1D (aged 41±16y, 64% female), average bolus frequency was 5.2±2.5/day, with 37% of users bolusing 3-4.99 times/day. Increased bolus frequency was correlated with improved GMI (Figure), decreased percentage of readings <54mg/dL, and increased percentage of readings 70-180mg/dL. The percentage of readings 70-180mg/dL increased from 39% with infrequent bolusing (<3/day) to 55% with frequent bolusing (≥8/day), while the percentage of readings <54mg/dL decreased from 2.8% to 1.7%.

Conclusions

Higher bolus frequency was associated with better glycemic control as measured by GMI and percentage of readings in target range in a large cohort of adults with T1D using the Omnipod System in this real-world observational study.

Background and Aims / Part 1

The first decade of continuous glucose monitoring (CGM) was relatively slow due to challenges with accuracy, burden, and cost. However, as we are now about half-way through our second decade of CGM with much higher usage, it is fair to ask the question and speculate who will be using SMBG five years from now?

.

Methods / Part 2

In the United States, in late 2017 Medicare, the payer for the elderly and disabled, started CGM coverage for individuals receiving multiple injections of insulin. The Medicare approval occurred shortly after the Freestyle Libre and Dexcom G5 launch, and thus this improved technology could be widely by those who previously couldn’t afford it. By late 2019 our type 1 population in our academic clinic was well over 70% CGM use. Our type 2 population, with over 80% using insulin, is over 25% CGM use (and no SMBG use)

Results / Part 3

However, our clinic is not the rule either in the US or elsewhere. We still have many payers in the US who will not pay for CGM and in fact are stingy with SMBG coverage. Worldwide the increase of diabetes is outpacing what we can afford. In the recent “IDF Atlas” it is predicted that in Africa there will be a 143% increase in diabetes prevalence. Currently in low-income countries there are approximately 60 million adults with diabetes where even obtaining insulin is a struggle. In these countries, it is not realistic to think CGM will become standard when currently SMBG is often difficult to obtain.

Conclusions / Part 4

SMBG is not going away in the next 5 years, or for that matter the next 10 years. For most people in the world, expense is the primary reason CGM can’t be a reality and SMBG, even occasional, will be the only way to monitor diabetes.

Background and Aims / Part 1

It has become common place for us to compare various patient registries in terms of A1C, rates of DKA, and severe hypoglycemia. We were curious to better understand glucometrics from CGM in the University of Washington Diabetes Care Center.

Methods / Part 2

From March 2017 to November 2019 we reviewed 247 patients with type 1 diabetes using CGM.

Results / Part 3

Average age was 52 years, average duration of diabetes was 28 years, half were women, and mean A1C was 7.2%. Sixty percent used Dexcom, 29% used Medtronic, and 11% used the Abbott Libre. Mean glucose was 163 mg/dL with a percent coefficient of variation (CV) of 34.8%. TIR for the population was 60.4% with Time Below Range (TBR) of 3.5%. GMI was 7.2%. Using mean glucose to predict GMI, 34% and 7% of measured A1C values were at least 0.5% and 1% discordant respectively.

Conclusions / Part 4

It is concluded that current consensus guidelines for TIR are reasonable. Percent CV below or equal to 33% is possible in T1D despite arguments this target should be below 36%. Finally, the GMI equation closely approximates our population’s CGM data.