Copyright 2017 American Medical Association. All rights reserved.
Effect of Continuous Glucose Monitoring on Glycemic Control
in Adults With Type 1 Diabete s Using Insulin Injections
The DIAMOND Randomized Clinical Trial
Roy W. Beck, MD, PhD; Tonya Riddlesworth, PhD; Katrina Ruedy, MSPH; Andrew Ahmann, MD;
Richard Bergenstal, MD; Stacie Haller, RD, LD, CDE; Craig Kollman, PhD; Davida Kruger, MSN, APN-BC;
Janet B. McGill, MD; William Polonsky, PhD; Elena Toschi, MD; Howard Wolpert, MD; David Price, MD;
for the DIAMOND Study Group
IMPORTANCE
Previous clinical trials showing the benefit of continuous glucose monitoring
(CGM) in the management of type 1 diabetes predominantly have included adults using
insulin pumps, even though the majority of adults with type 1 diabetes administer insulin
by injection.
OBJECTIVE To determine the effectiveness of CGM in adults with type 1 diabetes treated with
insulin injections.
DESIGN, SETTING, AND PARTICIPANTS Randomized clinical trial conducted be tween October
2014 and May 2016 at 24 endocrinology practices in the United States that included 158
adults with type 1 diabetes who were using multiple daily insulin injections and had
hemoglobin A
1c
(HbA
1c
) levels of 7.5% to 9.9%.
INTERVENTIONS Random assignment 2:1 to CGM (n = 105) or usual care (control group; n = 53).
MAIN OUTCOMES AND MEASURES Primary outcome measure was the difference in change
in central-laboratory–measured HbA
1c
level from baseline to 24 weeks. There were
18 secondary or exploratory end points, of which 15 are reported in this article, including
duration of hypoglycemia at less than 70 mg/dL, measured with CGM for 7 days at 12
and 24 weeks.
RESULTS Among the 158 randomized participants (mean age, 48 years [SD, 13]; 44% women;
mean baseline HbA
1c
level, 8.6% [SD, 0.6%]; and median diabetes duration, 19 years
[interquartile range, 10-31 years]), 155 (98%) completed the study. In the CGM group, 93%
used CGM 6 d/wk or more in month 6. Mean HbA
1c
reduction from baseline was 1.1% at 12
weeks and 1.0% at 24 weeks in the CGM group and 0.5% and 0.4%, respectively, in the
control group (repeated-measures model P < .001). At 24 weeks, the adjusted
treatment-group difference in mean change in HbA
1c
level from baseline was –0.6% (95% CI,
–0.8% to –0.3%; P < .001). Median duration of hypoglycemia at less than <70 mg/dL was
43 min/d (IQR, 27-69) in the CGM group vs 80 min/d (IQR, 36-111) in the control group
(P = .002). Severe hypoglycemia events occurred in 2 participants in each group.
CONCLUSIONS AND RELEVANCE Among adults with type 1 diabetes who used multiple daily
insulin injections, the use of CGM compared with usual care resulted in a greater decrease in
HbA
1c
level during 24 weeks. Further research is needed to assess longer-term effectiveness,
as well as clinical outcomes and adverse effects.
TRIAL REGISTRATION clinicaltrials.gov Identifier: NCT02282397
JAMA. 2017;317(4):371-378. doi:10.1001/jama.2016.19975
Editorial page 363
Related article page 379
Supplemental content
CME Quiz at
jamanetworkcme.com and
CME Questions page 436
Author Affiliations: Jaeb Center for
Health Research, Tampa, Florida
(Beck, Riddlesworth, Ruedy,
Kollman); Oregon Health & Science
University, Portland (Ahmann);
Park Nicollet Institute, International
Diabetes Center, St Louis Park,
Minnesota (Bergenstal); Diabetes &
Glandular Disease Clinic, San Antonio,
Texas (Haller); Division of
Endocrinology, Henry Ford Medical
Center, Detroit, Michigan (Kruger);
Washington University in St Louis,
St Louis, Missouri (McGill); Behavioral
Diabetes Institute, San Diego,
California (Polonsky); Joslin Diabetes
Center, Boston, Massachusetts
(Toschi, Wolpert); Dexcom Inc,
San Diego, California (Price).
Group Information: The DIAMOND
Study Group members are listed
at the end of this article.
Corresponding Author: Roy W. Beck,
MD, PhD, Jaeb Center for Health
Research, 15310 Amberly Dr, Ste 350,
Tampa, FL 33647 (rbeck@jaeb.org).
Research
JAMA | Original Investigation
(Reprinted) 371
Copyright 2017 American Medical Association. All rights reserved.
Downloaded From: https://jamanetwork.com/ on 08/26/2022
Copyright 2017 American Medical Association. All rights reserved.
O
nly approximately 30% of individuals with type 1 dia-
betes meet the American Diabetes Association goal of
hemoglobin A
1c
(HbA
1c
) level of 7.5% (58 mmol/mol)
for children (<18 years) and 7.0% (53 mmol/mol) for adults
(≥18 years),
1
indicating the need for better approaches to dia-
betes management. Continuous glucose monitoring (CGM)
with glucose measurements as often as every 5 minutes, plus
low and high glucose level alerts and glucose trend informa-
tion, has the capability of better informing diabetes manage-
ment decisions than blood glucose meter testing performed
several times a day. Randomized clinical trials have demon-
strated the benefit of CGM in adults with type 1 diabetes, but
not consistently in children, to improve glycemic control as
measured by HbA
1c
level and to reduce hypoglycemia.
2-6
These
previous trials have either completely or predominantly in-
cluded insulin pump users,
2,4,5
although the majority of adults
with type 1 diabetes deliver insulin via injections.
7,8
Only a small proportion of individuals with type 1 diabe-
tes who inject insulin use CGM, although the limited avail-
able observational data suggest that the glycemic benefit may
be comparable to that for pump users. In T1D Exchange reg-
istry 2015 data, mean HbA
1c
level in the 410 adult insulin in-
jecters using CGM was similar to that in 2316 pump users using
CGM (7.6% vs 7.7%, respectively) and lower than mean HbA
1c
level in the 6222 injection users not using CGM (7.6% vs 8.8%;
P < .001).
9
Whether individuals receiving insulin injections would be
willing to regularly wear CGM sensors and would derive glyce-
mic benefitsfrom CGM needs investigation. Accordingly, this ran-
domized multicenter clinical trial was conducted to evaluate the
effect of CGM in adults with type 1 diabetes who have elevated
HbA
1c
levels and use multiple daily injections of insulin.
Methods
The trial was conducted at 24 endocrinology practices in the
United States (19 community-based and 5 academic centers).
The protocol and Health Insurance Portability and Account-
ability Act–compliant informed consent forms were ap-
proved by institutional review boards (central commercial
board for 17 sites and local boards for the other 7 sites). Written
informed consent was obtained from each participant. The pro-
tocol is provided online and the statistical analysis plan is avail-
able in Supplement 1.
Study Participants
Major eligibility criteria included age 25 years or older, diagno-
sis of type 1 diabetes treated for at least 1 year with multiple daily
insulin injections, central laboratory–measured HbA
1c
level of
7.5% to 10.0%, no home use of a personal CGM device in the 3
months before the trial, and a negative pregnancy test for wom-
en of childbearing potential (eTable1 in Supplement 2 has a com-
plete listing of the inclusion and exclusion criteria).
Synopsis of Study De sign
Each participant was required to complete a 2-week preran-
domization phase using a CGM system that was configured to
record glucose concentrations not visible to the participant
(referred to as a “blinded” CGM). Eligibility required that the
blinded CGM be worn on at least 85% of possible days, the
CGM be calibrated at least 2 times per day, and blood glucose
meter testing (with a study-provided meter and test strips) be
performed at least 3 times daily. Fourteen participants did
not meet these criteria and did not continue into the random-
ized trial (Figure 1). One participant had a sudden death dur-
ing the prerandomization phase.
On the study website, after verification of eligibility
from data entered, each participant was assigned randomly
from a computer-generated sequence to either the CGM or con-
trol group in a 2:1 ratio, with a permuted block design (block
sizes of 3 and 6) stratified by HbA
1c
level (<8.5% and ≥8.5%).
A 2:1 randomization was used rather than 1:1 to provide a larger
sample size for a separate follow-on randomized trial assess-
ing glycemic benefits of initiating pump therapy in CGM
users using insulin injections.
Participants in the CGM group were provided with a CGM
system (Dexcom G4 Platinum CGM System with an enhanced
algorithm, software 505, Dexcom Inc) that measured glucose
concentrations from interstitial fluid in the range of 40 to
400 mg/dL every 5 minutes for up to 7 days. Participants in
both groups were provided with a Bayer Contour Next USB
meter and test strips. The CGM group was instructed to use
the CGM daily, calibrate the CGM twice daily, and verify the
CGM glucose concentration with the blood glucose meter
before injecting insulin (as per the regulatory labeling of the
device at the time the trial was conducted). General guide-
lines were provided to participants about using CGM, and
individualized recommendations were made by their clini-
cian about incorporating CGM trend information into their
diabetes management. The control group was asked to per-
form home blood glucose monitoring at least 4 times daily.
Participants in both groups were provided general diabetes
management education, and clinicians were encouraged to
review downloaded glucose data at each visit to inform treat-
ment recommendations, which were at clinician discretion
and not prescriptive in the protocol. eTable 2 in Supplement 2
describes the participant education as well as guidelines for
clinicians. CGM guidelines for participants are included in
Supplement 1.
Key Points
Question For adults with type 1 diabetes who are using multiple
daily insulin injections, does continuous glucose monitoring
improve hemoglobin A
1c
(HbA
1c
) levels compared with
self-monitored blood glucose management?
Findings In a randomized clinical trial of 158 adults with type 1
diabetes, there was a signif icantly greater decrease in HbA
1c
level
during 24 weeks with continuous glucose monitoring vs usual care
(–1.0% vs –0.4%).
Meaning Continuous glucose monitoring resulted in better
glycemic control compared with usual care, but further research
is needed to assess clinical outcomes, as well as effectiveness,
in a typical clinical population.
Research Original Investigation Glucose Monitoring and Glycemic Control via Insulin Injections
372 JAMA January 24/31, 2017 Volume 317, Number 4 (Reprinted) jama.com
Copyright 2017 American Medical Association. All rights reserved.
Downloaded From: https://jamanetwork.com/ on 08/26/2022
Copyright 2017 American Medical Association. All rights reserved.
Follow-up visits for both treatment groups occurred
after 4, 12, and 24 weeks. The CGM group had an additional
visit 1 week after randomization. The control group had 2
additional visits 1 week before the 12- and 24-week visits, at
which a CGM sensor in blinded mode was inserted to collect
glucose data for 1 week. Telephone contacts for both groups
occurred 2 and 3 weeks after randomization.
Hemoglobin A
1c
level was measured at baseline, 12
weeks, and 24 weeks at the Northwest Lipid Research Labo-
ratories, University of Washington, Seattle, with the Diabetes
Control and Complications Trial standardized analyzer
(TOSOH, Biosciences Inc).
Outcomes
The primary outcome was change in the central laboratory–
measured HbA
1c
level. Prespecified secondary outcomes
included percentage of participants with HbA
1c
level less than
7.0%; CGM-measured time in range (70-180 mg/dL), duration
of hypoglycemia (<70 mg/dL, <60 mg/dL, and <50 mg/dL),
duration of hyperglycemia (>180 mg/dL, >250 mg/dL, and
>300 mg/dL), and glucose variability (coefficient of varia-
tion); change in hypoglycemia unawareness
10
; and change in
frequency of blood glucose meter testing (longitudinal
changes in blood glucose meter testing were not assessed).
Prespecified exploratory outcomes included CGM-measured
mean glucose concentration and the following binary HbA
1c
outcomes to assist in translation of the primary HbA
1c
analy-
sis to a participant level: HbA
1c
level less than 7.5% and rela-
tive HbA
1c
reduction greater than or equal to 10%. Post hoc
outcomes included HbA
1c
reduction of 1% or more, HbA
1c
level less than 7.0% or reduction of 1% or more, CGM-
measured area above the curve 70 mg/dL and area under
the curve 180 mg/dL, change in insulin dose, and change in
body weight.
Satisfaction with CGM was assessed by completion at 24
weeks of the CGM Satisfaction Survey (44 items on a 1-5
Likert scale, with the computed score representing the
mean of the 44 items and subscales of benefits and lack of
hassles).
11
Quality-of-life and health economic outcomes
will be reported in separate articles.
Safety outcomes included severe hypoglycemia (defined
as an event that required assistance from another person
to administer carbohydrate, glucagon, or other resuscitative
actions), diabetic ketoacidosis, and serious adverse events re-
gardless of causality.
Statistical Methods
A sample size of 147 for the 2:1 randomization was calculated
to have 90% power to detect a difference in mean HbA
1c
level between treatment groups, assuming a population dif-
ference of 0.4%, standard deviation of the 24-week values of
0.7 adjusted for the correlation between baseline and
24-week values (based on data from the Juvenile Diabetes
Research Foundation CGM randomized trial
5
), and a 2-sided
α level of .05. Sample size initially was increased to 169 to
account for potential loss to follow-up. When it was recog-
nized by the coordinating center that the trial completion
rate was higher than anticipated, the recruitment goal was
changed to a minimum of 150, with the approval of the steer-
ing committee and the sponsor.
Analyses followed the intent-to-treat principle. The fol-
lowing change was made from the protocol and statistical
analysis plan before the data lock: the primary analysis was
a treatment group comparison of the change in HbA
1c
level
from baseline to 24 weeks, adjusted for baseline HbA
1c
level and clinical site as a random effect, in a repeated-
measures linear model in the protocol and with analysis of
covariance in the statistical analysis plan; both are reported
in this article. Confounding was assessed by repeating the
analysis, including potential confounding variables as covari-
ates. The Rubin method was used to impute for missing
data.
12
Exploratory analyses were conducted to assess for
interaction between the treatment effect on the change in
HbA
1c
level from baseline to 24 weeks and baseline factors by
including interaction terms in analysis of covariance models.
The following changes were made from the protocol and sta-
tistical analysis plan during the peer-review process: in post
hoc analyses, binary HbA
1c
outcomes were evaluated with
propensity scores
13
instead of logistic regression, adjusted for
baseline HbA
1c
level and clinical site; and for secondary,
exploratory, and post hoc analyses, 99% CIs instead of 95%
CIs are reported.
For CGM outcomes, treatment group comparisons using
the CGM data collected in each group for 7 days at 12 and 24
weeks were made with analysis of covariance models based
on ranks using van der Waerden scores if the metric was
Figure 1. Flowchart of Continuous Glucose Monitoring Study Completion
186 Patients enrolled
28 Excluded during
prerandomization phase
14 Did not meet run-in
eligibility for continuous
glucose monitoring and/or
blood glucose meter
1 Died
8 Requested to withdraw
5 Had HbA
1c
<7.5% or >10%
158 Randomized
105 Randomized to receive continuous
glucose monitoring system
53 Randomized to receive usual
care (control)
53 Completed the study3 Discontinued study
1 Lost to follow up
1 Site withdrew participant
1 Participant requested to
withdraw
102 Completed the study
2 Completed study but
discontinued continuous
glucose monitoring
53 Included in primary analysis105 Included in primary analysis
4 Imputation used for HbA
1c
value
All enrolled participants started the run-in phase; 28 did not proceed to
randomization for the reasons indicated in the figure. The number eligible for
screening who did not sign the informed consent form was not recorded.
Glucose Monitoring and Glycemic Control via Insulin Injections Original Investigation Research
jama.com (Reprinted) JAMA January 24/31, 2017 Volume 317, Number 4 373
Copyright 2017 American Medical Association. All rights reserved.
Downloaded From: https://jamanetwork.com/ on 08/26/2022
Copyright 2017 American Medical Association. All rights reserved.
skewed, adjusted for the corresponding baseline value, base-
line HbA
1c
level, and clinical site as a random effect. Similar
analyses were performed separately for daytime and night-
time. Frequency of blood glucose monitoring was compared
between groups with an analysis of covariance model, ad-
justed for the baseline frequency and clinical site as a ran-
dom effect.
Statistical methods for other analyses are described in table
footnotes. Standard deviations are reported for means and in-
terquartile ranges (IQRs) for medians where applicable. Re-
ported point estimates are unadjusted unless otherwise noted.
Analyses were conducted with SAS version 9.4. All P values
are 2 sided. P < .05 was considered significant for the pri-
mary analysis and P < .01 for all other analyses to account for
multiple comparisons (with 99% CIs accordingly provided).
SI Unit Conversions
Throughout, to convert HbA
1c
to the SI units of mmol/mol, mul-
tiply the HbA
1c
percentage value × 10.93 and subtract 23.5 from
the product. For example, an HbA
1c
value of 7.0% corre-
sponds to 53 mmol/mol. To convert glucose to mmol/L, mul-
tiply the values × 0.0555.
Results
Between October 2014 and December 2015, 158 participants
were assigned to the CGM group (n = 105) or control group
(n = 53). Mean age was 48 years (SD, 13) (range, 26-73 years,
with 34 participants [22%] ≥60 years); 44% were women. Me-
dian diabetes duration was 19 years (IQR, 10-31 years), and
mean baseline HbA
1c
level was 8.6% (SD, 0.6%; range, 7.5%-
9.9%). Participant characteristics according to randomized
group are shown in Table 1.
The 24-week primary study outcome visit was com-
pleted by 102 participants (97%) in the CGM group and all 53
(100%) in the control group (Figure 1). Overall visit comple-
tion was 99% and 98%, respectively. Three participants in the
CGM group (4 total visits) and 3 in the control group (3 total
visits) had additional visits, not required in the protocol, for
diabetes management.
Among the 102 participants in the CGM group who com-
pleted the trial, median CGM use was 7.0 d/wk (IQR, 7.0-7.0)
at 4, 12, and 24 weeks; only 2 (2%) discontinued CGM before
the 24-week visit. During month 6 (weeks 21-24), CGM use was
6 or more d/wk for 93% of the 102 participants (eTable 3 in
Supplement 2). No participant in the control group initiated
unblinded CGM use before the primary outcome.
According to meter downloads, mean blood glucose self-
monitoring was 5.1 tests per day (SD, 1.8) in the CGM group and
5.1 tests per day (SD, 1.4) in the control group during the base-
line period of blinded CGM wear and 3.6 tests per day (SD, 1.6)
and 4.6 tests per day (SD, 1.6), respectively, at 24 weeks
(adjusted mean difference for the change, –1.0; 99% CI, –1.7 to
–0.4; P < .001).
Glycemic Control and Other Outcomes
Primary Outcome
Mean reduction in HbA
1c
level from baseline was 1.1% at
12 weeks and 1.0% at 24 weeks in the CGM group and 0.5% and
0.4%, respectively, in the control group (primary analysis re-
peated-measures P < .001). At 24 weeks, the adjusted treat-
ment group difference in mean change in HbA
1c
level was –0.6%
(95% CI, –0.8% to –0.3%; P < .001) (Table 2). For each treat-
ment group, baseline and 24-week HbA
1c
values for each
Table 1. Baseline Participant Characteristics
Group, No. (%)
CGM
(n = 105)
Control
(n = 53)
Age, y
25-<45 53 (50) 16 (30)
45-<60 32 (30) 23 (43)
≥60 20 (19) 14 (26)
Mean (SD) [range] 46 (14)
[26-72]
51 (11)
[26-73]
Diabetes duration, median (IQR), y 19 (9-29) 19 (11-35)
Female sex 47 (45) 23 (43)
Highest education
a
<Bachelor’s degree 47 (47) 22 (43)
Bachelor's degree 43 (43) 19 (37)
Graduate degree 10 (10) 10 (20)
BMI, mean (SD) 28 (6) 27 (5)
Weight, mean (SD), kg 84 (20) 81 (18)
HbA
1c
,%
7.5-<8.5 47 (45) 24 (45)
8.5-≤9.9 58 (55) 29 (55)
Mean (SD) [range] 8.6 (0.7)
[7.5-9.9]
8.6 (0.6)
[7.5-9.9]
Self-reported No. of self-monitoring
blood glucose tests per day,
mean (SD)
3.9 (1.3) 4.1 (1.6)
Event in previous 12 mo
≥1 Severe hypoglycemia 8 (8) 9 (17)
≥1 Diabetic ketoacidosis 1 (<1) 1 (2)
Use of noninsulin glucose-lowering
medication
8 (8) 4 (8)
Total daily insulin dose, median (IQR),
U/kg/d
0.7 (0.5-0.9) 0.6 (0.5-0.9)
No. of long-acting insulin injections
per day
1 78 (74) 34 (64)
2 26 (25) 19 (36)
3 1 (<1) 0
No. of rapid-acting insulin injections
per day
2 0 1 (2)
3 71 (68) 32 (60)
4 23 (22) 15 (28)
≥5 11 (10) 5 (9)
CGM use previously 17 (16) 9 (17)
Abbreviations: BMI, body mass index (calculated as weight in kilograms divided
by height in meters squared); CGM, continuous glucose monitoring; HbA
1c
,
hemoglobin A
1c
; IQR, interquartile range.
SI Conversions: to convert HbA
1c
to the SI units of mmol/mol, multiply the HbA
1c
percentage value × 10.93 and subtract 23.5 from the product.
a
Education data missing for 5 in the CGM group and 2 in the control group.
Research Original Investigation Glucose Monitoring and Glycemic Control via Insulin Injections
374 JAMA January 24/31, 2017 Volume 317, Number 4 (Reprinted) jama.com
Copyright 2017 American Medical Association. All rights reserved.
Downloaded From: https://jamanetwork.com/ on 08/26/2022
Copyright 2017 American Medical Association. All rights reserved.
participant are shown in Figure 2A, and the cumulative dis-
tribution of the 24-week HbA
1c
values is shown in Figure 2B.
Secondary, Exploratory, and Post Hoc HbA
1c
Outcomes
The greater HbA
1c
improvement in the CGM group also was re-
flected in multiple participant-level secondary, exploratory, and
post hoc HbA
1c
outcomes (Table 2). There was no significant
interaction of the effect of treatment on 24-week HbA
1c
level
according to baseline HbA
1c
, age, education level, or type of
site (eTable 4 in Supplement 2).
Secondary and Exploratory CGM Outcomes
As secondary outcomes, CGM metrics for time in the range of
70 to 180 mg/dL, hyperglycemia, hypoglycemia, and glycemic
Figure 2. Hemoglobin A
1c
Values at Baseline and 24 Weeks, by Group
HbA
1c
at 24 Weeks, %
11
10
9
8
7
6
5
5 109
Baseline HbA
1c
, %
876
HbA
1c
at baseline and 24 weeks
A
CGM (n
=
102)
Control (n
=
53)
100
90
80
70
60
10
0
20
30
40
50
≤5 ≤10≤9
Cumulative Distribution
HbA
1c
at 24 Weeks, %
≤8≤7≤6
Cumulative distribution of HbA
1c
at 24 weeks
B
CGM (n
=
102)
Control (n
=
53)
A, Scatterplot of 24-week hemoglobin A
1c
(HbA
1c
) levels by baseline HbA
1c
level.
The horizontal line at 7.0% represents the American Diabetes Association HbA
1c
goal for adults with type 1 diabetes. Points below the diagonal line represent
cases in which the 24-week HbA
1c
level was lower than the baseline HbA
1c
level,
points above the diagonal line represent cases in which the 24-week HbA
1c
level
was higher than the baseline HbA
1c
level, and points on the diagonal line
represent cases in which the 24-week and baseline HbA
1c
values were the same.
B, Cumulative distribution of 24-week HbA
1c
values. For any given 24-week
HbA
1c
level, the percentage of cases in each treatment group with an HbA
1c
value at that level or lower can be determined from the figure. To convert HbA
1c
to the SI units of mmol/mol, multiply the HbA
1c
percentage value × 10.93 and
subtract 23.5 from the product.
Table 2. Primary Outcome and Hemoglobin A
1c
Outcomes at 12 and 24 Weeks
a
12 Weeks 24 Weeks
P Value
c,d
CGM Group
(n = 103)
Control
Group
(n = 52)
CGM Group
(n = 105)
b
Control
Group
(n = 53) Between-Group Difference
c,d
Primary outcome, mean (SD), % Mean adjusted difference, % (95% CI)
HbA
1c
7.6 (0.7) 8.1 (0.7) 7.7 (0.8) 8.2 (0.8)
Change in HbA
1c
from baseline −1.1 (0.7) −0.5 (0.7) −1.0 (0.8) −0.4 (0.7) −0.6 (−0.8 to −0.3) <.001
Prespecified secondary outcome, No. (%) Mean adjusted difference, % (99% CI)
HbA
1c
<7.0% 14 (14) 2 (4) 18 (18) 2 (4) 15 (0 to 30) .01
Prespecified exploratory outcomes, No. (%)
HbA
1c
<7.5% 49 (48) 6 (12) 39 (38) 6 (11) 31 (12 to 51) <.001
Relative reduction in HbA
1c
≥10% 62 (60) 12 (23) 58 (57) 10 (19) 37 (16 to 58) <.001
Post hoc outcomes, No. (%)
Reduction in HbA
1c
≥1% 55 (53) 12 (23) 53 (52) 10 (19) 33 (11 to 54) <.001
Reduction in HbA
1c
≥1% or HbA
1c
<7.0% 57 (55) 12 (23) 53 (52) 11 (21) 31 (9 to 52) <.001
Abbreviations: CGM, continuous glucose monitoring; HbA
1c
, hemoglobin A
1c
.
SI Conversion: to convert HbA
1c
to the SI units of mmol/mol, multiply the HbA
1c
percentage value × 10.93 and subtract 23.5 from the product.
a
Mean baseline HbA
1c
level was 8.6% in each group. For all analyses,
missing HbA
1c
values in which the central laboratory value was missing
but the local laboratory value was known were imputed with a regression line
based on the site’s local HbA
1c
measurements (CGM/control: 1/0 at 12 weeks;
1/0 at 24 weeks).
b
For the 24-week primary outcome only, the Rubin method was used to impute
missing HbA
1c
values when both the central and local laboratory values were
missing (3 in the CGM group and 0 in the control group). For the secondary,
exploratory, and post hoc analyses, n = 102.
c
For the primary analysis, treatment group comparisons were made with
analysis of covariance models, adjusted for baseline HbA
1c
level and clinical
site as a random effect. Model residuals were verified to have an approximate
normal distribution.
d
For the secondary, exploratory, and post hoc outcome s, treatment group
comparisons were made with propensity scores, adjusted for baseline HbA
1c
level and clinical site. P < .01 was considered significant to account for multiple
comparisons (with 99% CIs accordingly provided).
Glucose Monitoring and Glycemic Control via Insulin Injections Original Investigation Research
jama.com (Reprinted) JAMA January 24/31, 2017 Volume 317, Number 4 375
Copyright 2017 American Medical Association. All rights reserved.
Downloaded From: https://jamanetwork.com/ on 08/26/2022