DIABETES TECHNOLOGY & THERAPEUTICS

Volume 20, Number 10, 2018

ª Mary Ann Liebert, Inc.
DOI: 10.1089/dia.2018.0202

BRIEF REPORT

Retrospective Analysis of 3-Month Real-World Glucose

Data After the MiniMed 670G System Commercial Launch
Downloaded by Wegner Health Science Information Center/ University of South Dakota multisite from www.liebertpub.com at 09/03/18. For personal use only.

Michael P. Stone, MS, Pratik Agrawal, MS, Xiaoxiao Chen, PhD, Margaret Liu, BS,
John Shin, PhD, MBA, Toni L. Cordero, PhD, and Francine R. Kaufman, MD

Abstract
Real-world data from the first 3141 patients who completed 3 months of SmartGuard Auto Mode-enabled
MiniMed 670G system use during the MiniMed 670G System Commercial Launch are reported. CareLink
system data uploaded by real-world patients in the Commercial Launch from March 17, 2017 to December 31,
2017 were deidentified and analyzed. Comparisons of overall and night (10:00 PM–07:00 AM) time spent
below, within, and above target glucose range (TIR) (70–180 mg/dL) between the baseline Manual Mode and
closed-loop Auto Mode periods were made. These were evaluated alongside data from the 124 patients (aged 14–
75 years) who completed the 3-month MiniMed 670G system pivotal trial (NCT 2463097), from June 2, 2015 to
March 7, 2016. Real-world patients used Auto Mode a median 80.8% of the time (19 h and 24 min of the day).
The overall mean of time spent in TIR was 66.0% during baseline Manual Mode versus 73.3% during Auto Mode
(P < 0.001); the mean percentage of sensor glucose values <70 mg/dL was 2.7% versus 2.1% (P < 0.001); and that
>180 mg/dL was 31.4% versus 24.6% (P < 0.001). The nighttime and early morning (03:00 AM–06:00 AM) TIR
during Auto Mode was greater than that during baseline Manual Mode (nighttime: 77.2% vs. 67.4% [P < 0.001],
early morning: 70.9% vs. 84.6% [P < 0.001]). Similar differences between Manual Mode and Auto Mode TIR
were observed across different age groups. A slight increase in total insulin delivered was also observed. Con-
sistent with improved glycemic control demonstrated in the pivotal trial, analysis of CareLink system data from
>3000 real-world patients who completed 3 months of Auto Mode-enabled MiniMed 670G system use demon-
strated increased TIR and decreased time below and above TIR compared with baseline. These improved clinical
outcomes were observed across a broad age range of patients with type 1 diabetes.

Keywords: Real-world, Type 1 diabetes, Closed-loop system, Time in range

Introduction days, there were no episodes of severe hypoglycemia, dia-

betic ketoacidosis, or serious device-related adverse events.

T he Medtronic MiniMed 670G system with Smart-

Guard technology consists of the MiniMed 670G in-
sulin pump, the Guardian Sensor 3 continuous glucose
monitoring system, and the CareLink data management
system. The MiniMed 670G system is the first hybrid closed-
loop system that automatically adjusts basal insulin delivery
every 5 min based on sensor glucose (SG) data. The results of
the MiniMed 670G system pivotal trial (NCT 2463097) in
adolescents and adults with type 1 diabetes, which assessed 2
weeks of baseline data during Manual Mode and 3 months of
study phase data in which the SmartGuard Auto Mode feature
was enabled, showed that the system was safe and able to
improve glycemic control.1,2 During the *12,000 patient
From baseline to the end of the study phase, mean glycated
hemoglobin dropped from 7.4% to 6.9% (P < 0.001), the
percentage of time spent in the target glucose range (TIR) of
71–180 mg/dL increased (from 66.7% to 72.2% [P < 0.001]),
and the percentage of time spent in hyperglycemia and hy-
poglycemia was reduced (from 27.4% to 24.5% [P < 0.001]
and from 5.9% to 3.3% [P < 0.001], respectively).
The utility of real-world data as a source for clinical evi-
dence,3,4 and in the area of device therapy management,4 has
grown. The CareLink data management system allows up-
loads of insulin pump and glucose sensor data for ongoing
diabetes therapy management and retrospective analyses. In
this report, glycemic data from the first *3000 real-world

Medtronic, Northridge, California.

1
 2 STONE ET AL.
patients using the MiniMed 670G system were evaluated and
compared with data from the 124 adolescent and adult pa-
tients who completed the MiniMed 670G system pivotal trial.
Methods
Deidentified CareLink system data from real-world pa-
tients (n = 3141) who uploaded from March 17, 2017 to De-
cember 31, 2017 and who completed ‡3 months of MiniMed
670G system Auto Mode use were retrospectively analyzed.
The first upload to the CareLink system occurred a mean of
Downloaded by Wegner Health Science Information Center/ University of South Dakota multisite from www.liebertpub.com at 09/03/18. For personal use only.
15 days after Manual Mode start. A mean of eight uploads
occurred during the Auto Mode-enabled period. The 24-h and
nighttime (10:00 PM–07:00 AM) system use; mean SG val-
ues; mean percentage of SG values <50 mg/dL, <54 mg/dL,
<70 mg/dL, 70–180 mg/dL (TIR), >180 mg/dL, >250 mg/dL,
and >300 mg/dL; mean total insulin delivered; and mean
percentage of basal insulin delivered were determined. The
early morning (03:00 AM–06:00 AM) TIR was evaluated
across patients aged 7–13 years (n = 105), 14–21 years
(n = 244), 22–60 years (n = 2066), and ‡60 years (n = 649).
The system use and SG data from the 124 participants (n = 30
adolescents and n = 94 adults) who completed the MiniMed
670G system pivotal trial are presented for comparison.1 All
analyses conducted for the pivotal trial were considered ex-
ploratory. A 2-week run-in phase with the system in Manual
Mode served as baseline and was followed by a 3-month
study phase with Auto Mode enabled. In this study, analyses
were performed with either a Wilcoxon signed-rank test or
paired t-test where indicated. Age range-specific data were
evaluated based on the patient-reported age range recorded in
the CareLink data management system.
Results
The 24-h and nighttime system use in all patients
within each cohort
The Manual Mode and Auto Mode data (i.e., mean of
system use, median of Auto Mode use, overall mean of SG
values, mean of total insulin delivered, and mean percentage
of SG values in glucose ranges) for the pivotal trial cohort and
the real-world cohort for the 24-h and nighttime periods are
shown in Table 1. For the real-world cohort (total n = 3141
patients), there were 45,273 patient days/nights in Manual
Mode and 283,348 patient days/nights after Auto Mode was
enabled. For the pivotal trial cohort (total n = 124 patients),
there were 2277 and 12,291 patient days/nights, respectively.
For the real-world cohort 24-h and nighttime periods, there
was an increase in TIR and a decrease in the percentage of SG
values in the hypoglycemic (<70 mg/dL) and hyperglycemic
(>180 mg/dL) ranges during Auto Mode compared with
Manual Mode (Table 1). The overall mean SG values during
Manual Mode and Auto Mode were comparable with the
values seen in the pivotal trial cohort. During the early
morning hours of 03:00 AM to 06:00 AM, the real-world
cohort TIR during Manual Mode was 70.9%, whereas that
during Auto Mode was 84.6% (P < 0.001). For the pivotal
trial cohort, the early morning TIR during Manual Mode was
69.1% and that during Auto Mode was 82.2% (P < 0.001).
The total daily dose of insulin delivered from Manual
Mode to Auto Mode for the real-world patients and those in
the pivotal trial increased, for both the 24-h and nighttime
periods (Table 1). The percentage of the total daily dose of
insulin delivered as basal versus bolus, from Manual Mode
(49.1%) to Auto Mode (48.1%), for the real-world cohort 24-
h period decreased by 1.0%. From Manual Mode (92.4%) to
Auto Mode (92.6%) for the nighttime period, it increased
only slightly (by 0.2%). For the pivotal trial cohort 24-h
period, insulin delivered as basal versus bolus from Manual
Mode (53.0%) to Auto Mode (46.7%) decreased by 6.3%,
and for the nighttime period it decreased by 1.1%, from
Manual Mode (77.3%) to Auto Mode (76.2%).
Glycemic control in the real-world cohort,
by age groups
Patients aged 7–13 years in the real-world cohort had a TIR
of 56% during Manual Mode and a TIR of 67% during Auto
Mode; patients aged 14–21 years had a Manual Mode TIR of
57% and an Auto Mode TIR of 65%; those aged 22–60 years
had a Manual Mode TIR of 66% and an Auto Mode TIR of
74%; and for those aged >60 years, the Manual Mode TIR was
71%, whereas the TIR during Auto Mode was 77%. All age
groups experienced a reduction in the overall mean SG values
from Manual Mode to Auto Mode: 178 – 26 mg/dL to 161 –
16 mg/dL for 7–13 years; 174 – 30 mg/dL to 164 – 21 mg/dL
for 14–21 years; 158 – 27 mg/dL to 151 – 14 mg/dL for 22–60
years; and 154 – 21 mg/dL to 148 – 11 mg/dL for ‡60 years.
All age groups had a reduced percentage of SG values
>180 mg/dL: 42.2%–31.1% for 7–13 years; 40.6%–33.1%
for 14–21 years; 31.0%–24.2% for 22–60 years; and 27.2%–
21.4% for ‡60 years. For the percentage of SG values
<70 mg/dL, those between 7 and 13 years of age had 1.4%
during Manual Mode and 1.9% during Auto Mode, those
14–21 years of age had 2.5% and 2.2%, respectively; those
22–60 years of age had 2.9% and 2.2%, respectively; and
those ‡60 years of age had 2.3% and 1.9%, respectively. All
of the aforementioned changes in SG values >180 mg/dL and
<70 mg/dL, from the initial Manual Mode period to the end of
the Auto Mode-enabled period, were significant.
Discussion
The pivotal trial of the Medtronic MiniMed 670G system
with SmartGuard technology showed that the system was safe
and that it improved overall, nighttime, and early morning
glycemia in 124 participants for a 3-month duration.1 In this
study, the real-world data from the first 3141 patients who
used the system, as the approved commercial product, showed
similar benefits with Auto Mode use.
In real-world patients aged ‡14 years, there was a reduction
in 24-h and nighttime SG values in the hypoglycemic ranges.
Although those aged 7–13 years displayed a significant in-
crease in overall (but not nighttime) SG values <70 mg/dL,
this group had the lowest percentage of hypoglycemic SG
values at baseline and were not expected to show a decrease in
hypoglycemia during Auto Mode. There was no significant
increase in time (overall or night) spent at lower levels of hy-
poglycemia (i.e., <54 mg/dL or <50 mg/dL) for this group. A
limitation of this retrospective analysis is that the timing and
duration of hypoglycemic and hyperglycemic events were not
assessed.
The best overall glycemic results during Auto Mode were
observed in patients ‡60 years of age, whereas the youngest
cohort had the greatest reduction in mean SG values and SG
 Downloaded by Wegner Health Science Information Center/ University of South Dakota multisite from www.liebertpub.com at 09/03/18. For personal use only.

Table 1. Glycemic Outcomes with the MiniMed 670G System During Pivotal Trial and Real-World Use
Pivotal trial cohort Real-world cohort
24-h Nighttime 24-h Night-time
Run-in phase Study phase Run-in phase Study phase Manual Mode Auto Mode Manual Mode Auto Mode
(2 weeks) (3 months) P (2 weeks) (3 months) P (*2 weeks) (3 months) P (*2 weeks) (3 months) P
Patients, n 124 3141
Median Auto Mode use, % — 87.2 — — 84.8 — — 80.8 — — 79.9 —
Mean – SD of SG, mg/dL 150 – 23 151 – 14 0.2667 149 – 25 147 – 14 0.7341 159 – 27 152 – 15 <0.001 157 – 29 148 – 15 <0.001
Mean – SD of total 47.5 – 22.7 50.9 – 26.7 <0.001 12.6 – 6.9 13.3 – 8.3 0.009 49.8 – 27.6 51.9 – 29.3 <0.001 9.2 – 5.0 9.5 – 6.1 <0.001

3
insulin delivered, U/day
Mean percentage of time spent in SG ranges, mg/dL
<50 0.9 0.5 <0.001 1.0 0.6 <0.001 0.4 0.3 <0.001 0.4 0.3 <0.001
<54 1.3 0.8 <0.001 1.5 0.9 <0.001 0.6 0.5 <0.001 0.6 0.5 <0.001
<70 5.5 3.0 <0.001 6.0 2.9 <0.001 2.7 2.1 <0.001 2.6 1.9 <0.001
70–180a 67.1 72.4 <0.001b 67.2 75.5 <0.001b 66.0 73.3 <0.001 67.4 77.2 <0.001
>180 27.4 24.5 <0.001b 26.8 21.6 <0.001 31.4 24.6 <0.001 30.0 20.9 <0.001
>250 6.9 5.6 0.007 6.6 4.9 0.010 8.1 5.4 <0.001 7.2 4.3 <0.001
>300 2.3 1.7 0.184 2.1 1.4 0.750 2.5 1.6 <0.001 2.1 1.2 <0.001
Nighttime = 10:00 PM to 07:00 AM.
a
The target glucose range analyzed for the pivotal trial was 71–180 mg/dL1.
b
Paired t-test.
SD, standard deviation; SG, sensor glucose.
 4 STONE ET AL.
values in hyperglycemic ranges. Improvement in 24-h and
nighttime glycemic metrics, including reduced hypoglyce-
mic and hyperglycemic exposure, was seen with a minimal
increase in the total daily dose of insulin delivered.
Glycemic outcomes, such as TIR and percentage of SG
values in hypoglycemic and hyperglycemic ranges, are in-
valuable parameters to measure and report in the area of
innovative type 1 diabetes management.5,6 To this important
point, analysis of CareLink system data from the first 3141
patients using the MiniMed 670G system confirm that this
advanced insulin delivery therapy can reduce real-world low
Downloaded by Wegner Health Science Information Center/ University of South Dakota multisite from www.liebertpub.com at 09/03/18. For personal use only.
and high glucose levels and improve real-world glycemia in
pediatric and adult patients with diabetes.
Author Disclosure Statement
All authors are employees of Medtronic (Northridge, CA).
Medtronic funded the Hybrid Closed-Loop Pivotal Trial
in Type 1 Diabetes study (NCT02463097). The Medtronic
MiniMed 670G system is currently approved for use in the
United States for patients aged ‡7 years with type 1 diabetes.
References
1. Bergenstal RM, Garg S, Weinzimer SA, et al. Safety of a
hybrid closed-loop insulin delivery system in patients with
type 1 diabetes. JAMA 2016;316:1407–1408.
2. Garg SK, Weinzimer SA, Tamborlane WV, et al. Glucose
outcomes with the in-home use of a hybrid closed-loop in-
sulin delivery system in adolescents and adults with type 1
diabetes. Diabetes Technol Ther 2017;19:155–163.
3. Sherman RE, Anderson SA, Dal Pan GJ, et al. Real-world
evidence—what is it and what can it tell us? N Engl J Med
2016;375:2293–2297.
4. Tarricone R, Boscolo PR, Armeni P. What type of clinical
evidence is needed to assess medical devices? Eur Respir
Rev 2016;25:259–265.
5. Rodbard D. Continuous glucose monitoring: a review of
recent studies demonstrating improved glycemic outcomes.
Diabetes Technol Ther 2017;19:S25–S37.
6. Wright LA, Hirsch IB. Metrics beyond hemoglobin A1C in
diabetes management: time in range, hypoglycemia,
and other parameters. Diabetes Technol Ther 2017;19:S16–
S26.
Address correspondence to:
Toni L. Cordero, PhD
Medtronic
18000 Devonshire Street
Northridge, CA 91325
E-mail: toni.l.cordero@medtronic.com