Diabetes & Metabolism 44 (2018) 61–72

Available online at

ScienceDirect
www.sciencedirect.com

Position Statement

Practical implementation, education and interpretation guidelines for

continuous glucose monitoring: A French position statement
S. Borot a,*, P.Y. Benhamou b, C. Atlan c, E. Bismuth d, E. Bonnemaison e, B. Catargi f,
G. Charpentier g, A. Farret h, N. Filhol i, S. Franc j,k, D. Gouet l, B. Guerci m, I. Guilhem n,
C. Guillot o, N. Jeandidier p, M. Joubert q, V. Melki r, E. Merlen s, A. Penfornis t, S. Picard u,
E. Renard h, Y. Reznik p, J.P. Riveline t, S. Rudoni u, P. Schaepelynck i, A. Sola-Gazagnes v,
N. Tubiana-Rufi d, O. Verier-Mine w, H. Hanaire q, Société francophone du diabète (SFD),
Société française d’endocrinologie (SFE). Évaluation dans le diabète des implants actifs
Group (EVADIAC)
a
Department of Endocrinology, Nutrition and Diabetes, Besançon University Hospital and Franche-Comté University,
25030 Besançon cedex, France
b
Department of Diabetology, pôle DigiDune, Grenoble University Hospital, Grenoble Alpes University, 38700 La Tronche, France
c
Department of Endocrinology, Luxembourg Hospital, 1210 Luxembourg,Luxembourg
d
Department of Pediatric Endocrinology and Diabetology, Robert-Debré Hospital, AP–HP, 75019 Paris, France
e
Department of Pediatric Medicine, Tours University Hospital, 37044 Tours cedex, France
f
Department of Endocrinology and Diabetes, Bordeaux University Hospital, 33000 Bordeaux, France
g
Center for Study and Research for Improvement of the Treatment of Diabetes (CERITD), 91058 Evry cedex, France
h
Department of Endocrinology and UMR CNRS, Montpellier University Hospital and University of Montpellier,
34090 Montpellier, France
i
Department of Diabetology, Marseille University Hospital, 13005 Marseille, France
j
Department of Diabetology, Sud-Francilien Hospital, 91110 Corbeil-Essonnes, France
k
Department of Diabetology, La Rochelle General Hospital, 17000 La Rochelle, France
l
Department of Endocrinology, Diabetology, Metabolic Diseases and Nutrition, Nancy University Hospital,
54500 Vandœuvre-lès-Nancy, France
m
Department of Endocrinology and Diabetology, Rennes University Hospital, 35200 Rennes, France
n
Diabetes LAB, French Diabetes Federation, 75011 Paris, France
o
Department of Endocrinology and Diabetology, Strasbourg University Hospital, 67091 Strasbourg, France
p
Department of Endocrinology and Diabetology, Caen University Hospital, 14033 Caen, France
q
Department of Diabetology, Toulouse University Hospital, 31400 Toulouse, France
r
Department of Endocrinology, Lille University Hospital, 59000 Lille, France
s
Point Medical, Dijon Dijon, France
t
Department of Diabetes and Endocrinology, Lariboisière Hospital, University Paris 7, AP–HP, 75475 Paris, France
u
Department of Endocrinology, Diabetes and Metabolic Diseases, Dijon University Hospital, 21000 Dijon, France
v
Department of Diabetes, Cochin Hospital, AP–HP, 75014 Paris, France
w
Department of Diabetes, Valencienne General Hospital, 59300 Valencienne, France

Abbreviations: AGP, Average glucose profile; AJD, Association d’aide aux jeunes diabétiques (Young Diabetics Help Association); CGM, Continuous glucose monitoring;
CODEHG, Collège des diabétologues et endocrinologues des hôpitaux généraux (College of General Hospital Diabetologists and Endocrinologists); CNP-EDMM, Conseil
national professionnel d’endocrinologie, diabète et maladies métaboliques (National Professional Council of Endocrinology, Diabetes and Metabolic Diseases); CV, Coefficient
of variation; EVADIAC, Groupe d’évaluation dans le diabète des implants actifs (Evaluation Group of Active Implants in Diabetes); FFD, Fédération française des diabétiques
(French Diabetes Federation); FGM, Flash glucose monitoring; FSL, FreeStyle Libre; IG, Interstitial glucose; IQR, Interquartile range; PLGS, Predictive low-glucose suspend;
TLGS, Threshold low-glucose suspend; SFD, Société francophone du diabète (Francophone Society of Diabetes); SFE, Société française d’endocrinologie (French Society of
Endocrinology); SMBG, Self-monitoring blood glucose; T1D, Type 1 diabetes; T2D, Type 2 diabetes; TIR, Time in range.
* Corresponding author. Department of Endocrinology, Nutrition and Diabetes, hôpital Jean-Minjoz, 3, boulevard Fleming, 25030 Besancon cedex France.
E-mail address: sophie.borot@univ-fcomte.fr (S. Borot).

https://doi.org/10.1016/j.diabet.2017.10.009
1262-3636/ C 2017 Elsevier Masson SAS. All rights reserved.
62 S. Borot et al. / Diabetes & Metabolism 44 (2018) 61–72

A R T I C L E I N F O A B S T R A C T

Article history: The use by diabetes patients of real-time continuous interstitial glucose monitoring (CGM) or the
Received 21 July 2017 FreeStyle Libre1 (FSL) flash glucose monitoring (FGM) system is becoming widespread and has changed
Received in revised form 16 October 2017 diabetic practice. The working group bringing together a number of French experts has proposed the
Accepted 17 October 2017
present practical consensus. Training of professionals and patient education are crucial for the success of
Available online 11 November 2017
CGM. Also, institutional recommendations must pay particular attention to the indications for and
reimbursement of CGM devices in populations at risk of hypoglycaemia. The rules of good practice for
Keywords:
CGM are the precursors of those that need to be enacted, given the oncoming emergence of artificial
Continuous glucose monitoring
Flash glucose monitoring
pancreas devices. It is necessary to have software combining user-friendliness, multiplatform usage and
Guidelines average glucose profile (AGP) presentation, while integrating glucose and insulin data as well as events.
Patient education Expression of CGM data must strive for standardization that facilitates patient phenotyping and their
Subcutaneous insulin infusion follow-up, while integrating indicators of variability. The introduction of CGM involves a transformation
Type 1 diabetes of treatment support, rendering it longer and more complex as it also includes specific educational and
Type 2 diabetes technical dimensions. This complexity must be taken into account in discussions of organization of
diabetes care.
C 2017 Elsevier Masson SAS. All rights reserved.


Introduction
The use by diabetes patients of real-time continuous interstitial
glucose monitoring (CGM) or the FreeStyle Libre1 (FSL) flash
glucose monitoring (FGM) system is becoming more and more
widespread and has changed patient, caregiver and researcher
practices. Recommendations have been published recently for
CGM use and data-reporting in clinical trials [1]. The working
group bringing together a number of French experts [Conseil
national professionnel d’endocrinologie, Diabète et maladies
métaboliques (CNP-EDMM; National Professional Council of
Endocrinology, Diabetes and Metabolic Diseases), Société franco-
phone du diabète (SFD; Francophone Society of Diabetes), Société
française d’endocrinologie (SFE; French Society of Endocrinology),
Collège des diabétologues et endocrinologues des hôpitaux
généraux (CODEHG; College of General Hospital Diabetologists
and Endocrinologists), Groupe d’évaluation dans le diabète des
implants actifs (EVADIAC; Evaluation Group of Active Implants in
Diabetes), Fédération française des diabétiques (FFD; French
Diabetes Federation) and Association d’aide aux jeunes diabétiques
(AJD; Young Diabetics Help Association)] has proposed the present
consensus to assist professionals in integrating these new
technologies into their daily practice. Its main message is that
the training of professionals and patient education are crucial to
the success of CGM. The main recommendations of the working
group are summarized in Table 1.
What is measurement of interstitial glucose?
CGM/FGM devices are based on the semi-continuous measure-
ment of glucose in interstitial tissue. However, there is a
discrepancy between the displayed value of interstitial glucose
(IG) and that of capillary blood glucose due to the time delay of IG
equilibration relative to blood glucose as well as the delay with
measurements using subcutaneous electrodes due to converting
the electrical signal into glucose levels and displaying the results
on a screen [2,3]. Furthermore, the relationship between blood
glucose and IG is not just shifted in time, but is a more complex
pattern reflecting the dynamic profile of glycaemia, characterized
by a glucose lag (difference in glucose values in blood vs.
interstitial fluid at each time point) and a time lag (differences
in times when IG is equal to blood glucose) [4]. The delay is about
10 min with increased blood glucose, but can be shorter if it is
decreased (up to 6 min) [5]. The estimated time for FSL is
4.5  4.8 min [6].
As a result, the observed differences between capillary blood
glucose and IG are even greater when glycaemic variations are
extreme and rapid. Device trend arrows provide information on the
direction and speed of variations in IG levels ( 1–2 mg/dL/min for
the first level,  2–3 or > 2 mg/dL/min for the second level, and
> 3 mg/dL/min for the third level, depending on the CGM system).
The trends are generated from the slope of glucose values over the
previous 15 min and provide vital information for interpreting the
displayed values. The given information must be considered as
inseparable value/trend pairs for determining the action to be taken.
The different devices currently available
Table S1 (see supplementary data associated with this article
online) summarizes the main characteristics of the different
systems that are currently available.
CGM devices
Two types of devices provide real-time CGM:
 independent devices with sensor, transmitter and receiver:
 Dexcom G41 and G51 (Dexcom, San Diego, CA, USA),
 FreeStyle Navigator II1 (Abbott Laboratories, Chicago, IL,
USA),
 Guardian Connect1 (receiver is a smartphone or Apple iPod;
Medtronic, Minneapolis, MN, USA);
 devices with sensor and transmitter connected to a subcutane-
ous insulin pump, which acts as the receiver:
 Animas1 Vibe1 (Animas Corporation, West Chester, PA, USA),
 MiniMed 640G1 (Medtronic).
These systems need to be calibrated to capillary blood glucose
at least twice a day. The service life of the sensor is 5–7 days. They
are capable of producing alarms and some can automatically
suspend the basal rate of the pump when either hypoglycaemia
arises [threshold low-glucose suspend (TLGS) systems] or before it
happens [predictive low-glucose suspend (PLGS) systems]. Some
systems can remotely transmit data to a third party in real time
(Dexcom G5, Guardian Connect).
 S. Borot et al. / Diabetes & Metabolism 44 (2018) 61–72 63

Table 1
Summary of the French position statement on continuous glucose monitoring (CGM).
Indications FGM is considered an alternative to SMBG through the use of faster, easier, more frequent and more informative self-monitoring. Its use
is recommended as a replacement for SMBG in patients with T1D or T2D (adults and children) on intensified insulin therapy (pump or
multiple injections)

CGM systems are recommended in patients with T1D presenting with major hypoglycaemic problems (severe hypoglycaemia,
hypoglycaemia unawareness, phobia of hypoglycaemia). The working group suggests that particular attention be focused on populations at
risk of hypoglycaemia as per institutional recommendations concerning indications for and reimbursement of CGM devices

Real-time transmission of data should be considered in children and non-autonomous or isolated patients, and raises the issue of care
organization geared towards telemedicine

Patient education When starting CGM/FGM therapy, specific education should be provided to patients by caregivers trained in the use of these techniques
and follow-up as well as in therapeutic education, preferably with a multidisciplinary team

Initial education on FGM devices should include at least technical knowledge of sensor application, sensor reading and data entry, and
education on real-time decision-making according to values and trends

Initial education on CGM devices (FGM excluded) should include additional education on making calibrations and alarm management,
the settings of which should be defined with the patient

A 1-month trial is recommended before considering longer-term prescription of CGM

During follow-up, it is important to assess technical handling and potential problems such as wearing the system and its tolerability, data
management and impact on glycaemic control as well as patient satisfaction and quality of life (positive and negative points, alarm frequency)

Patient use of CGM/FGM
in real time
Values and trends should be used to adjust real-time decision-making:
– modulation of timing and level of insulin correction
– modulation of timing and quantity of carbohydrate intake to avoid or treat hypoglycaemia
– modulation of insulin pump basal rate
– increase in glucose checking frequency in specific situations (such as physical activity, unusual meals, driving)

Retrospective analysis Retrospective data analysis requires prior training of healthcare professionals (downloading and interpretation) and patients’ education

For data downloading, the working group recommends the use of secure and multiplatform software apps that provide an AGP and/or
standard daily profile and graphic representation of hypoglycaemic events, and allow IG and insulin data synchronization

It is recommended to prospectively add daily event notes (insulin dose and timing, food intake, physical activity to IG data for 1 or
2 weeks before the patient’s next visit

Analysis of patterns must take into account the patient’s personalized glycaemic goals, established with the agreement of both the
patient and healthcare professional

Excluding certain specific situations and with the goal of standardization in mind, the working group proposes the objective of > 60% of
time spent in the range of 70–180 mg/dL (3.9–10.0 mmol/L), with < 10% of time spent at < 70 mg/dL; a coefficient of variation > 36% is
considered excessive glucose variability

The structured data analysis must successively address the following points: is the level of compliance appropriate?; are basal insulin
doses correct?; is prandial insulin coverage effective?; are there recurrent hypo- or hyperglycaemic events?; are there indications of
device or data misuse?; the time and expertise of caregivers for downloading and analyzing data must also be considered

Specific characteristics The working group agrees with the international paediatric recommendations stating that the medical indications in children and
in children and adolescents adolescents are totally consistent with those of adults

FGM: flash glucose monitoring; SMBG: self-monitoring of blood glucose; T1D/T2D: type 1/type 2 diabetes; AGP: average glucose profile; IG: interstitital glucose.

FGM device: FreeStyle Libre
The Abbott FreeStyle Libre (FSL) has a sensor that can also be
used as a transmitter. Unlike CGM systems that automatically
communicate data to the receiver, the FSL receiver only displays a
value if the patient scans the sensor. The service life of the sensor is
14 days and no calibration is necessary. It is also possible to send
data remotely in real time (LibreLinkUp app).
The entire system (sensor, transmitter, receiver) must never be
exposed to X-rays (computed tomography, radiography) or
electromagnetic radiation (magnetic resonance imaging), but it
can withstand metal detectors.
Literature data
An alternative to self-monitoring of blood glucose (SMBG)
Glycaemic control of type 1 diabetes (T1D) retains priority over
recommended objectives [7], as HbA1c improvement is limited by
an increased risk of hypoglycaemia [8]. Although a relationship has
been established between daily SMBG and HbA1c levels [9], < 30%
of patients on a basal/bolus regimen perform the number of SMBG
that meets the recommendation of at least four measurements per
day [10]. In contrast, the IMPACT study [11], T1D patients using FSL
performed an average of 15.1  6.9 scans/day.
Metabolic and economic efficacy of CGM/FGM
Studies conducted during the first decade following the introduc-
tion of CGM have sampling biases related to the inclusion of both
patients with high HbA1c and those with hypoglycaemia in the same
study, resulting in neutralization of the results as regards HbA1c
improvement or reduction in hypoglycaemic events. CGM/FGM
efficacy should be based on glucose control (time spent in target range
and/or HbA1c), hypoglycaemic events, quality of life, patient
satisfaction and cost [12]. However, each parameter should be
integrated differently according to CGM/FGM indications (hypo-
glycaemia, high HbA1c, diabetes variability, diabetes-related dis-
tress. . .). Studies reported since 2012 has focused on either patients
64 S. Borot et al. / Diabetes & Metabolism 44 (2018) 61–72

with frequent hypoglycaemia and/or poor awareness of hypogly-
caemia, with a 40–50% reduction in time spent in hypoglycaemia
without modification [13] and even improvement in HbA1c of 0.2–
0.5% [14,15], or uncontrolled patients, with improvement in HbA1c of
0.43–0.6% and a reduction in hypoglycaemic events [16–19]. Efficacy
has been demonstrated in patients whether on pumps and receiving
multiple injections [16,18].
The FSL FGM system also demonstrated reductions of 38–43% in
time spent in hypoglycaemia without worsening of HbA1c in both
T1D [11] and type 2 diabetes (T2D) patients on basal/bolus regimens
[10]. The determinants of CGM efficacy are device usage time and
education on its use [20] and the number of ‘scans’ with FSL
[21]. CGM systems paired with a pump that offers automatic
suspensions (TLGS or PLGS) have shown benefits through reduction
of hypoglycaemia (severe and moderate) in patients with severe
hypoglycaemia or hypoglycaemia unawareness [13,22,23].
On the other hand, cost-effectiveness studies are complex, as
they need to consider in each population the cost of failure to
implement CGM/FGM, including the cost of emergency manage-
ment of severe hypoglycaemic events (emergency room visits,
hospitalizations, mortality, morbidity, work interruptions) and the
cost of decreased quality of life and long-term complications,
within a diversity of pricing systems [12]. A recent study supports a
favorable cost/efficacy ratio of CGM due to a 32% reduction in
hospitalizations for hypoglycaemia, a saving worth $54 million for
a T1D population of 46,500 patients [24].
Given the data in the literature, the working group considers
FGM an alternative to SMBG through the use of faster, easier, more
frequent and more informative self-monitoring. Its use is
recommended as a replacement for SMBG in patients with T1D
or T2D (adults and children) on intensified insulin therapy (pump
or multiple injections; Fig. 1).
CGM systems are recommended in patients with T1D
presenting with major hypoglycaemic problems (severe hypo-
glycaemia, hypoglycaemia unawareness, hypoglycaemia phobia).
The working group suggests that institutional recommendations
pay particular attention to populations at risk of hypoglycaemia in
terms of indications for and reimbursement of CGM devices
(Fig. 1). Also, real-time transmission of data should be considered
in children and in non-autonomous or isolated patients and raise
the issue of care organization geared towards telemedicine.
 patient responsiveness with regard to therapeutic adjustments;
 patient commitment: as the efficacy of CGM correlates with
sensor wearing time, patients must agree to use it permanently
after being informed of the related conditions and usage
constraints.
Initial therapeutic education
The use of CGM devices must be guided by structured
therapeutic education, which is essential for the acquisition of
technical skills as well as the capacity to manage glucose
information in both real time and retrospectively [19]. Such
education needs to be provided by caregivers trained in the use of
the techniques as well as in therapeutic education, preferably with
a multidisciplinary team, and should also aim to assess the
expectations, fears or difficulties of patients and their personal
experience of the use of these new tools.
Trial period and evaluation
A 1-month trial period is recommended, followed by an
evaluation to identify technical problems, compliance (wearing
time, number of scans), benefits and drawbacks, and the personal
experience of the patient, before considering longer-term CGM
prescription.
Initial technical education and patient follow-up
Organization of data collection (CGM/FGM)
For retrospective analysis of the data, it is important to train
patients in prospective data collection. An exhaustive collection
over a representational period (2 weeks) prior to consultation will
make the retrospective analysis more effective for treatment
adjustment. It will therefore be necessary to document meals
(time, quantity of carbohydrates, daily food journal, photos. . .),
insulin doses and timings; activities (hobbies, sports); medication
treatments (acetaminophen) [25]; and sleep position (side, back or
stomach) because of their possible interference with CGM
measurement. It is recommended that these parameters be entered
directly into the system to facilitate the retrospective analysis.

Initial therapeutic education

Calibrations (CGM only)

Prerequisites Special attention should be focused on teaching patients how to

perform calibrations during a period of glycaemic stability.
The following are necessary before the introduction of CGM: Patients must also be made aware of the possibility of erroneous
 good manual dexterity; values if calibration is performed poorly [26]. Evening calibration is
[(Fig._1)TD$IG] good cognitive ability;
 also recommended to avoid nocturnal ‘reminders’.

‘Hypoglycaemia issues’?
Frequent episodes that impact quality of life, severe hypoglycaemic events,
hypoglycaemia unawareness, phobia of hypoglycaemia

YES NO

CGM devices FGM devices (FreeStyle Libre®)

± automatic basal rate suspension (TLGS, PLGS)

Fig. 1. Algorithm for choosing continuous (CGM)/flash glucose monitoring (FGM) devices for diabetes patients on insulin pump therapy or multiple injections. TLGS/PLGS:
threshold/predictive low-glucose suspend.
 S. Borot et al. / Diabetes & Metabolism 44 (2018) 61–72
Alarm settings (CGM only)
It is essential to limit the number of activated alarms initially,
and perhaps even to not have any activated in the first place, to
avoid immediate rejection of the system. If an alarm does become
activated, patients need to know how to change the settings or
deactivate it. This requires specific educational assessment.
There are several alarm categories, including:
 ‘threshold’ alarms, which warn patients when the IG level
crosses the ‘hyper’ or ‘hypo’ threshold. The threshold for the
hypo alarm should be determined with the patient (50–90 mg/
dL) and the upper limit can also vary considerably according to
the patient (180–300 mg/dL). It is recommended that the ‘hyper’
alarm be set very high (> 350 mg/dL) to detect infusion
problems or forgotten boluses, as setting it at < 200 mg/dL is
liable to trigger a large number of warnings;
 ‘predictive’ alarms, which warn patients before IG levels reach
the predefined hypo- or hyperglycaemia threshold (after
establishing the time interval before the hypo- or hyperglycae-
mia threshold is reached). These take into account not only the
IG value, but also the trend, and they precede ‘threshold’ alarms,
allowing patients to act before hypoglycaemia manifests
(preventative glucose administration, temporary pump output).
Activation of the predictive ‘hyper’ alarm is not recommended;
 ‘trend’ or ‘speed’ alarms, which warn patients of an increase or
decrease in IG levels that are too rapid, regardless of value, as
reflected by the number of arrows or speed in mg/dL/min. These
alarms may sometimes ring to indicate an expected glycaemic
variation (correction of hyperglycaemia, glucose administration
following hypoglycaemia). An upward trend alert can indicate a
forgotten bolus; a downward trend alert can sometimes
forewarn of imminent sudden hypoglycaemia;
 ‘reminder’ alarms or ‘repetitive’ alerts, which repeat an alarm
based on a defined interval of time until the parameter is corrected.
The MiniMed Paradigm Veo1 and 640G1 (Medtronic) insulin
pumps paired with CGM have additional TLGS or PLGS alarms. The
former is accompanied by a strong audible warning for a threshold
that cannot be < 50 mg/dL, whereas the latter alarm can be activated
on discontinuation and/or resumption of the basal rate; most
patients prefer to delete these alarms and to only be warned if the
hypoglycaemic threshold is crossed despite basal rate discontinua-
tion. The recommended PLGS threshold is generally 60–65 mg/dL,
but can vary from 55 mg/dL to 70 mg/dL, depending on the patient.
Patient follow-up
It is important to assess technical handling and any potential
problems, wearing the system and its tolerability, patient
Table 2
Summary of real-time therapeutic decisions according to glucose values and trends.
IG # Stable
Below target Carb intake
Pump:  temporary basal rate Recheck in 15 min
On target Carb intake if ## or symptoms
Pump:  temporary basal rate
Recheck in 30–60 min
Above target Recheck in 1–2 h
Or give correction bolus
(unless ## or last bolus < 2 h)
IG: interstitial glucose; KT: catheter.
65
satisfaction and impact on quality of life, number of daily alarms
and effects on glycaemic control with the CGM system.
Use of CGM/FGM in real time
Compared with standard SMBG, the management of insulin
therapy is modified by the additional information (more values,
changing trends) offered by CGM. This results in more frequent
decision-making, leading to more rapid metabolic improvement
[11,18]. However, decision-making itself becomes more complex,
requiring therapeutic education to limit the risks of glycaemic
instability, anxiety and/or rejection of the technology.
Use of arrow trends in real time
Learning how to analyze IG levels and trends is essential for
good therapeutic decisions in real time and patients need to
integrate value/trend data pairs before taking action (Table 2).
Contribution of CGM/FGM to corrective short-acting insulin doses in
real time
An anticipatory correction factor taking into account arrow
trends (an increase of 10% or 20% with the appearance of single or
double arrows) has been proposed for rapid calculation of insulin
doses [27]. Others have proposed calculating the short-acting
insulin dose from the predicted glucose level at 30 min [4].
Contribution of CGM/FGM to management of hypoglycaemia in real
time
With hypoglycaemic symptoms, glucose administration is
indicated if the displayed value is low or normal with a downward
trend [28]. If the displayed value is low but with a trend upwards, it
is possible to wait and recheck, especially if there was prior glucose
administration.
In the absence of symptoms in patients on multiple daily
injections, notification of impending hypoglycaemia (predictive
hypoglycaemic alarm, down-pointing trend arrows with values
that are still normal) can help to decide whether to stop the
physical activity underway, implement controlled preventative
glucose administration or continue but with frequent monitoring
(every 5 min).
In the absence of symptoms and the IG value is not too low,
patients on pumps can choose to temporarily lower their basal rate
or turn it off for a mean duration of 30 min, which can be renewed
if necessary, while bearing in mind that each 30-min interruption
of infusion results in a delayed 30-mg/dL increase in blood glucose
[29,30]. It is important to remind patients who use basal rate
suspension, rather than temporary basal rate reduction, to restart
"
Carb intake if symptoms Carb intake if symptoms
Pump: temporary basal rate (and last carb intake > 15 min)
Recheck in 15–30 min Recheck
[TD$INLE] Recheck in 30–60 min
Correction bolus Correction bolus
(unless last bolus < 2 h) (unless last bolus < 2 h)
Recheck in 1–2 h Recheck in 1–2 h
Pump: ketone and KT control
(if fasting IG > 250 mg/dL)
66 S. Borot et al. / Diabetes & Metabolism 44 (2018) 61–72
the pump when values return to > 100 mg/dL with a trend arrow
that is at least stable. Stopping the basal rate for more than 2 h can
expose patients to hyperglycaemic rebound [23], which is often
slow and delayed. PLGS systems avoid having patients manually
manage flow rate interruptions and limit the risk of too-long basal
rate suspension.
CGM/FGM helps to manage the quantity of glucose adminis-
tration, which should be smaller when trend arrows point upwards
and glucose is taken to prevent hypoglycaemia, and when the basal
rate had been stopped for several minutes, spaced apart from a
bolus. Patients with PLGS need to take care not to take glucose
systematically whenever the pump preventatively stops and,
instead, wait for symptoms or the hypoalarm. This can be difficult
for patients who fear hypoglycaemias. Conversely, glucose
quantities can be larger when the IG value is low with a down-
pointing arrow.
It is reasonable to propose IG monitoring at least every 10 to
15 min as long as the IG is not > 100 mg/dL with a horizontal or
upward trend arrow.
CGM/FGM and driving
It is important to check the screen, receiver or scanner before
driving and, thereafter, at least every 2 h. European law only
authorizes driving [31] with very regular blood glucose monitor-
ing, which patients must be able to demonstrate if involved in a
vehicular accident. If IG is < 90 mg/dL (or higher if the trend is
clearly downwards), patients must either stop or postpone
departure and eat instead [32]. Similarly, if the level is high but
with a downward trend, it is advisable to wait before taking a
corrective bolus or take a smaller bolus.
Contribution of CGM/FGM to meal coverage in real time
Apart from certain situations that require particularly strict
glycaemic control, it is inadvisable to correct postprandial
hyperglycaemia within 1.5 h of the first prandial bolus to avoid
‘overbolusing’. If necessary, using bolus calculators, which take
into account on-board insulin, can limit secondary hypoglycaemia.
CGM can prove to be particularly helpful for irregular
mealtimes, as it can guide patients’ decisions to split boluses or
manage the potentially delayed postprandial hyperglycaemia.
When should the screen or scanner be consulted for making
a decision?
The frequency of IG value consultations varies greatly from
one patient to another as well as in the same patient, depending
on events (symptoms, trend arrows) and the system (CGM, FSL).
The working group considers it imperative to check when
symptoms are present, and at least before each carbohydrate
intake and/or insulin injection, and 2 h afterwards. It is also
recommended to check before and at least every hour during
physical activity.
Checks can be frequent (every 5 to 10 min) if the trend is
downwards with a normal–low IG value in the absence of glucose
intake. However, patients need to learn to not ‘overreact’ and to
recheck values to make sure there is a real trend that requires
action.
When should capillary blood glucose be checked?
Except for calibration, the need for or use of capillary
measurements should be discussed with the patient. According
to the current data, capillary blood glucose checks are not required
before any therapeutic action (glucose administration, correction,
flow rate change. . .) [19]. Nevertheless, patients need to be aware
that some sensors can be defective and that the FSL system is less
reliable during the first 24 h of use [6]. However, any discrepancy
between the value given by the system and clinical judgement
requires verification with capillary blood glucose.
Retrospective analysis and self-analysis of IG profiles
The retrospective analysis of data requires prior training of
healthcare professionals (downloading and interpretation) and
patient education [33] and is largely facilitated by collecting daily
events (insulin doses and timing, food, physical activity. . .) in real
time at least 1–2 weeks before consultation. Analysis of patterns
has to take into account the personalized glycaemic goals of the
patient, mutually established by both the patient and professional.
Structured analysis of the data must successively address the
following points:
 is the level of compliance appropriate?
 are basal insulin doses correctly regulated?
 is prandial insulin coverage effective?
 are there recurrent hypo- or hyperglycaemic events?
 are there any indications of misuse of the device or data?
How to download?
Dedicated programmes are used to harvest data. The stan-
dardized presentation proposed for the average glucose profile
(AGP) combines graphic and statistical information into a single
figure and provides three new insights into glucose control:
 medians of averaged glucose patterns reflect total exposures to
glucose;
 high fluctuating interquartile ratios and 10th–90th percentiles
between the different AGP timepoints reflect high between-day
glucose variability – in other words, a lack of synchrony in day-
to-day glucose patterns;
 fluctuations of AGP medians in the horizontal axis reflect within-
day variability, although high between-day glucose variability
should raise the question of whether or not within-day
variability has been correctly assessed. In most cases, high
between-day variability does not allow any firm conclusions
regarding within-day glucose variability, but it is highly likely
that, in such circumstances, within-day variability can be
dubbed ‘erratic’.
The working group supports the use of programmes providing
the AGP, as such an attempt at standardization can facilitate the
analysis and comparison of data. The aim is to have a report that
synchronizes IG data with daily events (doses of insulin,
carbohydrate intake, physical activity . . .); this calls for graphic
representations of hypoglycaemic events (number, cumulative
durations). Some software features (automatic analysis of patterns
with possible explanations for the hypo- or hyperglycaemic event,
analysis of postprandial trends) can reduce the time-consuming
nature of comprehensive data processing. The goal is to design
software apps that combine user-friendliness and multiplatform
usage (any CGM/FGM system, Mac/PC, Web, smartphones/tablets).
Who downloads and how often?
The impact of retrospective CGM data analysis by caregivers on
metabolic control has been the subject of several randomized
controlled studies, with equivocal results for HbA1c: no change in
adults [34–36]; negative in children and adolescents [37–39]; and
 S. Borot et al. / Diabetes & Metabolism 44 (2018) 61–72
favorable with HbA1c reductions of 0.3–0.4% [40,41]. A prospective
paediatric study suggested that therapeutic modifications related
to CGM favor a reduction in the incidence of hypoglycaemias
[42]. During pregnancy, retrospective analysis of IG data every 4–6
weeks in T1D and T2D patients led to reductions in HbA1c in the
third trimester and in macrosomia compared with a control group
followed without CGM [43]. This positive effect of retrospective
CGM was the result of education.
Concerning the retrospective analysis of data by patients, a
recent single-centre, cross-sectional study of 155 adults and
185 parents of T1D children, some of whom were CGM users,
focused on ‘frequent downloaders’ ( 4 downloads/year), showed
that this proactive behavior was associated with significantly
lower HbA1c, with a difference approaching 1% [44]. Downloading
of data by patients themselves, at their own pace, can be
encouraged if they have received the interpretation keys.
As for data transmission by patients, we must strive for the
implementation of a dedicated platform for telemedicine, thereby
enabling automated analysis, and especially management with a
diagnostic and educational approach. This is facilitated by
cooperative protocols between doctors and nurses, with task
delegation. Indeed, treatment adjustment according to IG values is
a complex task that is often difficult to master by many patients,
who would therefore benefit from support by telemedicine
approaches.
The working group acknowledges that the level of proof is still
too low to claim that the retrospective analysis of IG data directly
contributes to improving HbA1c or reducing the incidence of
hypoglycaemia. However, it is acknowledged that programmes for
downloading data play a facilitating role in the diagnostic task of
caregivers in the technical management of insulin therapy and a
motivational and educational role in patients.
Treatment compliance
System wearing time can be deemed optimal if it enables the set
therapeutic objectives to be achieved. The benefits for HbA1c are
also significantly correlated with sensor usage time [45]. However,
the EVADIAC sensor study [19] showed that the achievement of
[(Fig._2)TD$IG]
Fig. 2. Patient metabolic phenotyping. DKA: diabetic ketoacidosis; ER: emergency room (from Bergenstal et al. [46]).
67
metabolic benefit does not require permanent use of the system
but, instead, only a minimum use 40% of the time.
Metabolic phenotyping
The working group agrees with the recent attempts to
standardize glycaemic data expression [46], which is helpful for
categorizing patients and evaluating therapeutic impact. These
include:
 categorization of different glycaemic levels according to their
clinical impact, based on the spectrum summarized in Fig. 2;
 use of time spent in the glycaemic target zone [time in range
(TIR)]; the working group proposes a TIR objective of 60% at 70–
180 mg/dL (3.9–10.0 mmol/L), with < 10% of time spent at
< 70 mg/dL;
 estimation of glycaemic variability, which is usually based on
the dispersion of glucose values around the 24-h mean glucose
concentration [46,47].
From a statistical point of view, the standard deviation (SD) is
the most appropriate index for assessing glucose variability.
However, the SD is usually positively correlated with mean glucose
concentrations. As a consequence, the coefficient of variation (%CV)
for glucose, calculated using the formula ([SD of glucose]/[mean
glucose concentration])  100, appears to be the most reliable
index for quantifying glycaemic variability, as it is adjusted against
mean glucose concentration [48]. When the %CV is computed from
its mean over several consecutive days, it appears that the
threshold for separating ‘stable’ from ‘labile’ diabetes states can be
set at 36% [49]. However, it must be noted that the SD and %CV do
not have the same meaning when computed using means
calculated from data collected every day over several consecutive
days and when derived from AGP values. The two indices can be
referred to as the ‘mean daily %CV’ and ‘daily %CV by average’,
respectively. However, the latter, which is easily calculated using
the averaged glucose and related SD values, is automatically
indicated on the first page of CGM reports, and is usually lower and
less reliable than the former [50].
68 S. Borot et al. / Diabetes & Metabolism 44 (2018) 61–72
Contribution of CGM/FGM to adjustment of basal insulin dose
CGM/FGM provides a more precise evaluation of the basal
insulin dose effect. Its use is highly recommended during a fasting
test. It also helps to identify nocturnal episodes and the ‘dawn
phenomenon’ or early-morning hypoglycaemia and to make
treatment adjustments. It can identify when the pump is stopped
too frequently, but not when disconnections fail to stop the pump,
which can then only be suggested by gradual increases in blood
glucose. In the presence of high nocturnal variability, it is advisable
to consider low values (curve of the 10th percentile) before
increasing basal rates. The determination of basal needs during the
day and in the evening is more complex due to interference of the
prandial bolus effect. In this case, a short fast by skipping a single
meal can be useful.
In addition, CGM/FGM contributes by documenting the marked
influence of physical activity on basal insulin needs during and
after the activity and also facilitates the implementation of suitable
preventative measures [51].
Indeed, the availability of precise 24-h data can encourage
patients to programme different basal rates over 24 h. However,
due to the kinetics of subcutaneous insulin, there are no benefits to
multiplying the number of different basal rates and the time ranges
of pump basal rates should generally not be < 4 h.
The use of TLGS or PLGS raises the issue of adjusting basal rates
when repeated automatic stops occur within a given period. It
seems evident that the basal rate should be reduced when the
hypoglycaemic threshold is crossed despite stopping the pump
apart from a bolus. It has also been acknowledged that a
cumulative duration of > 3–4 h of overnight stopping observed
repeatedly indicates an overdose of basal insulin, which can create
instability through hyperglycaemic rebound.
Use for determining prandial insulin doses
It is difficult to obtain consistent postprandial glycaemic
control in daily practice, and a single postprandial value, taken at
any hour, is a very poor reflection of the glycaemic changes that
take place rapidly after ingestion of carbohydrates and injection
of prandial boluses. For a precise analysis, the working group
recommends that meals be tagged on the curve to facilitate
interpretation of the observed IG variations. When unloading
data, it is possible to configure it for the usual mealtimes, which
can make the automatic analysis more relevant, especially if
the patient has unconventional mealtimes (shift workers, for
example).
Once meals have been identified, it is important to focus on the
postprandial curves for each meal while attempting to identify a
reproducible tendency. Some programmes superimpose a pro-
posed pattern of curves from the beginning of the identified meal
and focus more on the postprandial period to be studied. In other
systems, the superimposed curves or AGP can rapidly display the
glycaemic trend during main meals, provided that the meals are
generally taken at the same time of day.
Three situations can be considered:
 there is a trend towards a postprandial hyperglycaemic gradient
(> 1 g/L), with few hypoglycaemic events. This suggests the
prandial insulin dose is probably insufficient, but it might also be
that the prandial bolus was given too late in relation to the start
of the meal or perhaps missed completely. With the AGP, the
recurrence of forgotten boluses can appear graphically, ranging
from a dome-shaped postprandial hyperglycaemia to significant
glycaemic variability. Substantial postprandial hyperglycaemia
followed by an IG return to within-target values, or even
hypoglycaemia within 3 h, should raise the possibility of a bolus
given after the meal or apart from it;
 there is a marked trend towards postprandial hypoglycaemia.
This suggests the prandial insulin dose is probably overestima-
ted;
 there is wide postprandial variability. Apart from repeated
forgetfulness or variable times of the bolus relative to meals, the
most probable explanation is poor carbohydrate- (and/or lipid-)
counting. In this case, it is helpful to work with a daily journal.
Contribution of retrospective analysis to correction of hypoglycaemia
Excessive hypoglycaemia correction is revealed by glycaemic
excursions beyond target values following a hypoglycaemic
episode. Some systems offer automatic analysis of events
preceding hyperglycaemia and can identify hypoglycaemia as
preceding a hyperglycaemic peak. CGM/FGM can promote
overcorrection of hypoglycaemia due to the time lag between
glucose intake and IG increase as visualized by patients, who
should be warned of this unavoidable time lag. In addition, the
possible anxiety-provoking effect of continually visualizing IG
results can induce inappropriate preventative glucose administra-
tion measures in patients fearful of hypoglycaemia, thereby
sometimes worsening the situation through prolonged interrup-
tion of basal rates [52].
Use of retrospective analysis for correction of hyperglycaemia
The use of CGM/FGM in real time can prompt patients to give
themselves corrective boluses more frequently because of the
graphic visualization of hyperglycaemic excursions. While this
action is advisable, such corrective boluses may sometimes be too
frequent or too close together, which may result in hypoglycaemia.
This tendency can be spotted by the frequency of boluses or by a
low basal/bolus ratio. Patient education must be improved by
advising the systematic use of a bolus calculator for any corrections
and/or consideration of the active insulin before any manual bolus
is delivered.
Similarly, visualization of several correction boluses with no
significant effects can lead to an increase in correction boluses
(decrease in sensitivity coefficient).
Analysis of patient experiences
Most patients on CGM experience an improvement in quality of
life and greater social and professional freedom. Some, however,
develop inappropriate behaviors, with overinvestment in or
dependency on the sensor, with an abnormally high number of
scans, unreasonable treatment adjustments (> 5 basal rate slots,
> 5 corrective boluses/day, frequent use of temporary basal rate)
or pathological anxiety over usage interruption of sensors
(manifested by repetitive capillary blood glucose tests to mimic
continuous information from a sensor). The presence of these
behaviors must be assessed and could raise doubts over
continuation of CGM. They also highlight the importance of
therapeutic education and patient support.
Identification and correction of inappropriate uses and
behaviors
Table 3 summarizes the causes to bear in mind when there are
reliability problems with results, frequent misuse or insulin
therapy effects on CGM/FGM.
 S. Borot et al. / Diabetes & Metabolism 44 (2018) 61–72 69

Table 3
Summary of technical issues and continuous glucose monitoring (CGM) misuse or inappropriate management of insulin therapy under CGM.

Problems Causes Actions to take

Reduced data accuracy and/or
IG/SMBG discrepancy
First 24 h of FreeStyle Libre (FSL) sensor [4] Start new sensor at bedtime, with caution as to value
accuracy on the first night
Calibration issues (CGM only): mishandling, calibration Pay attention to calibration timing/values when
delay [23], IG lability period, wrong SMBG value downloading data
Partial or complete sensor detachment Check sensor, check sensor application, adapt adhesion
support
Defective sensor Change sensor
Sensor life exceeded Follow manufacturer recommendations
Acetaminophen (overestimation by 20–30 mg/dL for Avoid acetaminophen, inform patients
7 h) [22]
No interference described with FSL
Vitamin C, aspirin
Pressure artefact (low IG value returns to normal after Avoid applying sensor to pressure areas (waist, sleeping
pressure is released) position)

Repeated alerts (CGM only) Unnecessary alarms Evaluate alert frequency based on downloads, patient’s
Premature thresholds point of view (risk of demotivation, lack of reactivity)

Poor glucose control Missed prandial boluses Analyze data with patient, try to understand precisely
Delayed prandial boluses how patient is dealing with the situation
Carbohydrate variability not covered by insulin
Excessive basal rate suspensions
Basal/bolus imbalance
Overbolusing (overinvestment)
Premature or excessive hypoglycaemia corrections

Anxiety Psychological dependence on sensor Evaluate psychological impact and quality of life related
to CGM/FGM use

IG: interstitial glucose; SMBG: self-monitoring of blood glucose; FGM: flash glucose monitoring; Cobelli et al. [4]; Bergenstal et al. [23]; Buckingham et al. [22].

Specific characteristics in children and adolescents
Specific studies
The results of a dozen randomized controlled trials conducted
in children (> 1000 children and adolescents with T1D) since
2006 show a remarkable effect on hypoglycaemic events [20],
particularly with the TLGS function [13]. The benefits for HbA1c are
significant, albeit slightly inferior to those obtained in adults. These
benefits remain particularly dependent on device wearing time
and disappear when its use is discontinued. The available data ‘in
real life’ from large cohorts of young patients highlight the need to
identify possible obstacles to the use of the technology [53]; this is
because only 4–6% of children use CGM in the long term [54] and,
of those, < 15% download data regularly. In this context, the
establishment of therapeutic educational and structured support
programmes for families seems mandatory.
Indications for CGM systems
The working group agrees with the international paediatric
recommendations [55,56] stating that the medical indications in
children and adolescents are totally consistent with those of
adults. The distinctiveness of paediatric cases lies in the reasons for
the implementation of such systems when system requests can
come from caregivers, and also patients and their families, thus
requiring an educational assessment that takes into account family
issues and the motivations of relatives and patients, which can
sometimes differ (a desire for alarms and remote data access vs. an
obligation to take better care of their diabetes, risk of alarms at
night or during social life, aesthetic aspects, fear of being
monitored. . .).
Choice of system in children
Certain system characteristics may be important, such as the
diameter and length of the electrode and guide needle and size and
weight of the device. Recent devices enabling data transmission to
parents in real time have clear safety benefits, but also raise the
issue of freedom of choice in older children as to whether or not to
share their data in real time.
Sensor insertion
As children and parents often fear insertion of the sensor, it is
important to prepare the family well and to prevent pain, using an
anaesthetic cream during initial insertions. The choice of site,
particularly in young children, depends on subcutaneous tissue
thickness, while prioritizing the upper buttocks. We recommend
supporting families during the first two or three insertions until
there is complete comfort, as that is an important factor for long-
term acceptability. Trained nurses can also be supportive during
initial insertions.
Setting alarm thresholds
Setting alarm thresholds depends on the context – age, diabetes
stability, insulin regimen, priority objectives – of the given patient.
For children who are not under parental care during the day, we
recommend not activating the hyperalert, which could create
difficulties for nannies or teachers.
The PLGS function has its limitations when the basal rate is low
(or even zero), which is commonly seen in paediatric cases,
especially in the morning [57]. Patients and their families need to
be warned of these limitations, but reassured that the hypoalert
will ring in case of system failure.
Real-time use in paediatric cases
Real-time use depends on the fact that children are away from
their parents during most of the day and lack enough autonomy to
react appropriately. With the hypoalarm, children are trained to
notify an adult informed by their parents as to the correct quantity
of carbohydrates to administer and how to read the screen to verify
70 S. Borot et al. / Diabetes & Metabolism 44 (2018) 61–72
the increase in glucose. For children using FSL, the scan may be
done by either the child or an adult trained by the parents to do a
scan if hypoglycaemia is suspected. In practice, few real-time
adjustments are made in young children during the day and new
tools for remote data transmission to parents in real time are
eagerly anticipated. In older children, however, these tools
require active authorization by the patient (to respect privacy)
and the course of action will have been chosen together with the
parents.
Self-analysis/retrospective analysis of patterns
Frequent self-analysis (weekly vs. infrequently) has shown an
improvement of 1% in HbA1c [58]. It is therefore important to train
patients/parents on how to download and analyze CGM data in
order to adjust their treatment and evaluate the actions to be
undertaken.
Conclusion
CGM in real time is a considerable step towards refining useful
tools for the daily management of diabetes. Maintained by
caregiver and patient training, to which the present work aims to
contribute, it allows an integrated mode of treatment that is a
precursor to future advances towards an artificial pancreas.
Indeed, we have already witnessed its first fruits with the
availability of insulin pumps connected to glucose sensors
capable of interrupting their output when hypoglycaemia occurs
or is anticipated. Such systems have already generated a change
in attitude towards hypoglycaemic episodes and what to do when
the situation arises. More recently, a regulation on the adminis-
tration of insulin in hyperglycaemia was authorized in the US.
Thus, little by little, we are moving closer to the artificial
pancreas. In addition to the technological advances made by
device manufacturers and algorithm engineers, clinical skills and
solid patient training now need to be added before taking the next
steps in this development.
The working group responsible for the present recommenda-
tions emphasizes certain promising aspects in the short- and
medium-term future of these technologies:
 institutional recommendations need to pay particular attention
to populations at hypoglycaemic risk as regards the indications
for and reimbursement of CGM devices;
 the rules of good practice for CGM, especially those related to
technical training in the use of sensors and knowledge of the
causes of measurement error, are precursors to those that need
to be enacted with the oncoming emergence of artificial
pancreas devices;
 the time required for complete processing of CGM data by
healthcare professionals is considerable. Consequently, it is
necessary to design software combining user-friendliness,
multiplatform usage and AGP presentation that also integrates
data on glucose and insulin concentrations and events;
 the expression of CGM data must strive for standardization to
facilitate the phenotyping of patient data and follow-up of their
evolution. Indicators of variability (%CV, interquartile range, TIR)
are all helpful in the evaluation of quality of treatment;
 the introduction of CGM, while improving the quality of
management for patients with diabetes, has also involved the
transformation of treatment support, rendering it longer and
more complex, as it also comprises specific educational and
technical dimensions. This complexity must be taken into
account in discussions of the organization of care in diabetology
and their evaluation.
Funding
This work has been supported by the SFD, which covered
scientific meeting travel expenses.
Disclosure of interest
S. Borot discloses congress invitations, honoraria and consul-
tancies from Abbott, Animas/Johnson & Johnson, Medtronic, Roche.
P.-Y. Benhamou discloses congress invitations, honoraria and
consultancies from Abbott, Animas/Johnson & Johnson, Dexcom,
Insulet, Medtronic, Roche.
C. Atlan discloses congress invitations, honoraria and consul-
tancies from Abbott and Medtronic.
E. Bonnemaison discloses congress invitations, honoraria and
consultancies from Medtronic.
B. Catargi discloses congress invitations, honoraria and
consultancies from Abbott.
G. Charpentier discloses congress invitations, honoraria and
consultancies from Abbott, Dexcom, Medtronic.
N. Filhol discloses congress invitations, honoraria and consul-
tancies from Roche.
S. Franc discloses congress invitations, honoraria and consul-
tancies from Animas/Johnson & Johnson, Roche.
D. Gouet discloses congress invitations, honoraria and consul-
tancies from Abbott.
B. Guerci discloses congress invitations, honoraria and consul-
tancies from Abbott, Dexcom, Medtronic, Roche.
C. Guillot discloses congress invitations, honoraria and consul-
tancies from Abbott, Dexcom.
M. Joubert discloses congress invitations, honoraria and
consultancies from Abbott, Medtronic.
A. Penfornis discloses congress invitations, honoraria and
consultancies from Abbott, Medtronic.
S. Picard discloses congress invitations, honoraria and consul-
tancies from Abbott, Animas/Johnson & Johnson, Medtronic.
E. Renard discloses congress invitations, honoraria and
consultancies from Abbott, Animas/Johnson & Johnson, Dexcom,
Insulet, Medtronic, Roche, Ypsomed, Cellnovo.
Y. Reznik discloses congress invitations, honoraria and consul-
tancies from Medtronic.
J.-P. Riveline discloses congress invitations, honoraria and
consultancies from Abbott, Animas/Johnson & Johnson, Dexcom,
Medtronic.
P. Schaepelynck discloses congress invitations, honoraria and
consultancies from Roche.
A. Sola-Gazagnes discloses congress invitations, honoraria and
consultancies from Abbott, Medtronic.
N. Tubiana discloses congress invitations, honoraria and
consultancies from Abbott, Insulet, Medtronic.
O. Verier-Mine discloses congress invitations, honoraria and
consultancies from Abbott.
H. Hanaire discloses congress invitations, honoraria and
consultancies from Abbott, Animas/Johnson & Johnson, Medtronic,
Roche.
A. Farret, I. Guilhem, N. Jeandidier, V. Melki, E. Merlen,
E. Bismuth and S. Rudoni declare that they have no competing
interest.
Appendix A. Supplementary data
Supplementary data (Table S1) associated with this article can
be found, in the online version, at http://www.sciencedirect.com
and http://dx.doi.org/10.1016/j.diabet.2017.10.009.
 S. Borot et al. / Diabetes & Metabolism 44 (2018) 61–72
References
[1] Schnell O, Barnard K, Bergenstal R, Bosi E, Garg S, Guerci B, et al. Role of
continuous glucose monitoring in clinical trials: recommendations on
reporting. Diabetes Technol Ther 2017;19:391–9. http://dx.doi.org/
10.1089/dia.2017.0054.
[2] Pleus S, Schoemaker M, Morgenstern K, Schmelzeisen-Redeker G, Haug C, Link
M, et al. Rate-of-change dependence of the performance of two CGM systems
during induced glucose swings. J Diabetes Sci Technol 2015;9:801–7. http://
dx.doi.org/10.1177/1932296815578716.
[3] Ward WK, Engle JM, Branigan D, El Youssef J, Massoud RG, Castle JR. The effect
of rising vs. falling glucose level on amperometric glucose sensor lag and
accuracy in type 1 diabetes. Diabet Med J Br Diabet Assoc 2012;29:1067–73.
http://dx.doi.org/10.1111/j.1464-5491.2011.03545.x.
[4] Cobelli C, Schiavon M, Dalla Man C, Basu A, Basu R. Interstitial fluid glucose is
not just a shifted-in-time but a distorted mirror of blood glucose: insight from
an in silico study. Diabetes Technol Ther 2016;18:505–11. http://dx.doi.org/
10.1089/dia.2016.0112.
[5] Monsod TP, Flanagan DE, Rife F, Saenz R, Caprio S, Sherwin RS, et al. Do sensor
glucose levels accurately predict plasma glucose concentrations during hypo-
glycemia and hyperinsulinemia? Diabetes Care 2002;25:889–93.
[6] Bailey T, Bode BW, Christiansen MP, Klaff LJ, Alva S. The performance and
usability of a factory-calibrated flash glucose monitoring system. Diabetes
Technol Ther 2015;17:787–94. http://dx.doi.org/10.1089/dia.2014.0378.
[7] Miller KM, Foster NC, Beck RW, Bergenstal RM, DuBose SN, DiMeglio LA, et al.
Current state of type 1 diabetes treatment in the U.S.: updated data from the
T1D Exchange clinic registry. Diabetes Care 2015;38:971–8. http://dx.doi.org/
10.2337/dc15-0078.
[8] The Diabetes Control Complications Trial Research Group. The effect of inten-
sive treatment of diabetes on the development and progression of long-term
complications in insulin-dependent diabetes mellitus. N Engl J Med
1993;329:977–86. http://dx.doi.org/10.1056/NEJM199309303291401.
[9] Miller KM, Beck RW, Bergenstal RM, Goland RS, Haller MJ, McGill JB, et al.
Evidence of a strong association between frequency of self-monitoring of
blood glucose and hemoglobin A1c levels in T1D exchange clinic registry
participants. Diabetes Care 2013;36:2009–14. http://dx.doi.org/10.2337/
dc12-1770.
[10] Haak T, Hanaire H, Ajjan R, Hermanns N, Riveline J-P, Rayman G. Flash glucose-
sensing technology as a replacement for blood glucose monitoring for the
management of insulin-treated type 2 diabetes: a multicenter, open-label
randomized controlled trial. Diabetes Ther Res Treat Educ Diabetes Relat
Disord 2017;8:55–73. http://dx.doi.org/10.1007/s13300-016-0223-6.
[11] Bolinder J, Antuna R, Geelhoed-Duijvestijn P, Kröger J, Weitgasser R. Novel
glucose-sensing technology and hypoglycaemia in type 1 diabetes: a multi-
centre, non-masked, randomized controlled trial. Lancet Lond Engl
2016;388:2254–63. http://dx.doi.org/10.1016/S0140-6736(16)31535-5.
[12] Rodbard D. Continuous glucose monitoring: a review of successes, challenges
and opportunities. Diabetes Technol Ther 2016;18(Suppl 2):S3–13. http://
dx.doi.org/10.1089/dia.2015.0417.
[13] Ly TT, Nicholas JA, Retterath A, Lim EM, Davis EA, Jones TW. Effect of sensor-
augmented insulin pump therapy and automated insulin suspension vs.
standard insulin pump therapy on hypoglycemia in patients with type 1 dia-
betes: a randomized clinical trial. JAMA 2013;310:1240–7. http://dx.doi.org/
10.1001/jama.2013.277818.
[14] Choudhary P, Ramasamy S, Green L, Gallen G, Pender S, Brackenridge A, et al.
Real-time continuous glucose monitoring significantly reduces severe hypo-
glycemia in hypoglycemia-unaware patients with type 1 diabetes. Diabetes
Care 2013;36:4160–2. http://dx.doi.org/10.2337/dc13-0939.
[15] Battelino T, Phillip M, Bratina N, Nimri R, Oskarsson P, Bolinder J. Effect of
continuous glucose monitoring on hypoglycemia in type 1 diabetes. Diabetes
Care 2011;34:795–800. http://dx.doi.org/10.2337/dc10-1989.
[16] Beck RW, Riddlesworth T, Ruedy K, Ahmann A, Bergenstal R, Haller S, et al.
Effect of continuous glucose monitoring on glycemic control in adults
with type 1 diabetes using insulin injections: the DIAMOND randomized
clinical trial. JAMA 2017;317:371–8. http://dx.doi.org/10.1001/jama.2016.
19975.
[17] Battelino T, Conget I, Olsen B, Schütz-Fuhrmann I, Hommel E, Hoogma R, et al.
The use and efficacy of continuous glucose monitoring in type 1 diabetes
treated with insulin pump therapy: a randomized controlled trial. Diabeto-
logia 2012;55:3155–62. http://dx.doi.org/10.1007/s00125-012-2708-9.
[18] Lind M, Polonsky W, Hirsch IB, Heise T, Bolinder J, Dahlqvist S, et al. Continuous
glucose monitoring vs. conventional therapy for glycemic control in adults
with type 1 diabetes treated with multiple daily insulin injections: the GOLD
randomized clinical trial. JAMA 2017;317:379–87. http://dx.doi.org/10.1001/
jama.2016.19976.
[19] Riveline J-P, Schaepelynck P, Chaillous L, Renard E, Sola-Gazagnes A, Penfornis
A, et al. Assessment of patient-led or physician-driven continuous glucose
monitoring in patients with poorly controlled type 1 diabetes using basal-
bolus insulin regimens: a 1-year multicenter study. Diabetes Care
2012;35:965–71. http://dx.doi.org/10.2337/dc11-2021.
[20] Benhamou P-Y, Catargi B, Delenne B, Guerci B, Hanaire H, Jeandidier N, et al.
Real-time continuous glucose monitoring (CGM) integrated into the treat-
ment of type 1 diabetes: consensus of experts from SFD. EVADIAC and SFE
Diabetes Metab 2012;38(Suppl. 4):S67–83. http://dx.doi.org/10.1016/
S1262-3636(12)71538-0.
71
[21] Dunn T, Xu D, Hayter G. Evidence of a strong association between frequency of
flash glucose monitoring and glucose control measures during real-world
usage. Diabetes Technol Ther 2017;19(Suppl. 1):A12.
[22] Buckingham BA, Raghinaru D, Cameron F, Bequette BW, Chase HP, Maahs DM,
et al. Predictive low-glucose insulin suspension reduces duration of nocturnal
hypoglycemia in children without increasing ketosis. Diabetes Care
2015;38:1197–204. http://dx.doi.org/10.2337/dc14-3053.
[23] Bergenstal RM, Klonoff DC, Garg SK, Bode BW, Meredith M, Slover RH, et al.
Threshold-based insulin-pump interruption for reduction of hypoglycemia. N
Engl J Med 2013;369:224–32. http://dx.doi.org/10.1056/NEJMoa1303576.
[24] Bronstone A, Graham C. The potential cost implications of averting severe
hypoglycemic events requiring hospitalization in high-risk adults with type
1 diabetes using real-time continuous glucose monitoring. J Diabetes Sci
Technol 2016;10:905–13. http://dx.doi.org/10.1177/1932296816633233.
[25] Basu A, Veettil S, Dyer R, Peyser T, Basu R. Direct evidence of acetaminophen
interference with subcutaneous glucose sensing in humans: a pilot study.
Diabetes Technol Ther 2016;18(Suppl. 2):S243–7. http://dx.doi.org/10.1089/
dia.2015.0410.
[26] Thomas F, Signal M, Harris DL, Weston PJ, Harding JE, Shaw GM, et al.
Continuous glucose monitoring in newborn infants: how do errors in calibra-
tion measurements affect detected hypoglycemia? J Diabetes Sci Technol
2014;8:543–50. http://dx.doi.org/10.1177/1932296814524857.
[27] Diabetes Research In Children Network (DirecNet) Study Group, Buckingham
B, Xing D, Weinzimer S, Fiallo-Scharer R, Kollman C, et al. Use of the DirecNet
applied treatment algorithm (DATA) for diabetes management with a real-
time continuous glucose monitor (the FreeStyle Navigator). Pediatr Diabetes
2008;9:142–7. http://dx.doi.org/10.1111/j.1399-5448.2007.00301.x.
[28] Jensen MH, Christensen TF, Tarnow L, Mahmoudi Z, Johansen MD, Hejlesen OK.
Professional continuous glucose monitoring in subjects with type 1 diabetes:
retrospective hypoglycemia detection. J Diabetes Sci Technol 2013;7:135–43.
[29] Scheen A, Castillo M, Jandrain B, Krzentowski G, Henrivaux P, Luyckx AS, et al.
Metabolic alterations after a two-hour nocturnal interruption of a continuous
subcutaneous insulin infusion. Diabetes Care 1984;7:338–42.
[30] Zisser H. Quantifying the impact of a short-interval interruption of insulin-
pump infusion sets on glycemic excursions. Diabetes Care 2008;31:238–9.
http://dx.doi.org/10.2337/dc07-1757.
[31] diabetes_and_driving_in_europe_final_1_en.pdf n.d.
[32] Graveling AJ, Frier B. Driving and diabetes: problems, licensing restrictions and
recommendations for safe driving. Clin Diab Endoc 2015. http://dx.doi.org/
10.1186/s40842-015-0007-3.
[33] Joubert M, Baillot-Rudoni S, Catargi B, Charpentier G, Esvant A, Franc S, et al.
Indication, organization, practical implementation and interpretation guide-
lines for retrospective CGM recording: a French position statement. Diabetes
Metab 2015;41:498–508. http://dx.doi.org/10.1016/j.diabet.2015.07.001.
[34] Chico A, Vidal-Rı́os P, Subirà M, Novials A. The continuous glucose monitoring
system is useful for detecting unrecognized hypoglycemias in patients with
type 1 and type 2 diabetes but is not better than frequent capillary glucose
measurements for improving metabolic control. Diabetes Care 2003;26:
1153–7.
[35] Tanenberg R, Bode B, Lane W, Levetan C, Mestman J, Harmel AP, et al. Use of the
continuous glucose monitoring system to guide therapy in patients with
insulin-treated diabetes: a randomized controlled trial. Mayo Clin Proc
2004;79:1521–6. http://dx.doi.org/10.4065/79.12.1521.
[36] Cooke D, Hurel SJ, Casbard A, Steed L, Walker S, Meredith S, et al. Randomized
controlled trial to assess the impact of continuous glucose monitoring on
HbA(1c) in insulin-treated diabetes (MITRE study). Diabet Med J Br Diabet
Assoc 2009;26:540–7. http://dx.doi.org/10.1111/j.1464-5491.2009.02723.x.
[37] Chase HP, Kim LM, Owen SL, MacKenzie TA, Klingensmith GJ, Murtfeldt R, et al.
Continuous subcutaneous glucose monitoring in children with type 1 diabetes.
Pediatrics 2001;107:222–6.
[38] Yates K, Hasnat Milton A, Dear K, Ambler G. Continuous glucose monitoring-
guided insulin adjustment in children and adolescents on near-physiological
insulin regimens: a randomized controlled trial. Diabetes Care 2006;29:1512–
7. http://dx.doi.org/10.2337/dc05-2315.
[39] Deiss D, Hartmann R, Schmidt J, Kordonouri O. Results of a randomized
controlled cross-over trial on the effect of continuous subcutaneous glucose
monitoring (CGMS) on glycaemic control in children and adolescents with
type 1 diabetes. Exp Clin Endocrinol Diabetes 2006;114:63–7. http://
dx.doi.org/10.1055/s-2006-923887.
[40] Ludvigsson J, Hanas R. Continuous subcutaneous glucose monitoring im-
proved metabolic control in pediatric patients with type 1 diabetes: a con-
trolled crossover study. Pediatrics 2003;111:933–8.
[41] Lagarde WH, Barrows FP, Davenport ML, Kang M, Guess HA, Calikoglu AS.
Continuous subcutaneous glucose monitoring in children with type 1 diabetes
mellitus: a single-blind, randomized, controlled trial. Pediatr Diabetes
2006;7:159–64. http://dx.doi.org/10.1111/j.1399-543X.2006.00162.x.
[42] Schiaffini R, Ciampalini P, Fierabracci A, Spera S, Borrelli P, Bottazzo GF, et al.
The continuous glucose monitoring system (CGMS) in type 1 diabetic children
is the way to reduce hypoglycemic risk. Diabetes Metab Res Rev 2002;18:324–
9. http://dx.doi.org/10.1002/dmrr.309.
[43] Murphy HR, Rayman G, Lewis K, Kelly S, Johal B, Duffield K, et al. Effectiveness
of continuous glucose monitoring in pregnant women with diabetes: random-
ized clinical trial. BMJ 2008;337:a1680.
[44] Wong JC, Neinstein AB, Spindler M, Adi S. A Minority of patients with type
1 diabetes routinely downloads and retrospectively reviews device data.
72 S. Borot et al. / Diabetes & Metabolism 44 (2018) 61–72
Diabetes Technol Ther 2015;17:555–62. http://dx.doi.org/10.1089/dia.
2014.0413.
[45] Pickup JC, Freeman SC, Sutton AJ. Glycaemic control in type 1 diabetes during
real time continuous glucose monitoring compared with self monitoring of
blood glucose: meta-analysis of randomized controlled trials using individual
patient data. BMJ 2011;343:d3805.
[46] Bergenstal RM, Ahmann AJ, Bailey T, Beck RW, Bissen J, Buckingham B, et al.
Recommendations for standardizing glucose reporting and analysis to opti-
mize clinical decision making in diabetes: the ambulatory glucose profile
(AGP). Diabetes Technol Ther 2013;15:198–211. http://dx.doi.org/10.1089/
dia.2013.0051.
[47] Rodbard D. Clinical interpretation of indices of quality of glycemic control and
glycemic variability. Postgrad Med 2011;123:107–18. http://dx.doi.org/
10.3810/pgm.2011.07.2310.
[48] Rodbard D. Hypo- and hyperglycemia in relation to the mean, standard
deviation, coefficient of variation and nature of the glucose distribution.
Diabetes Technol Ther 2012;14:868–76. http://dx.doi.org/10.1089/
dia.2012.0062.
[49] Monnier L, Colette C, Wojtusciszyn A, Dejager S, Renard E, Molinari N, et al.
Toward defining the threshold between low and high glucose variability
in diabetes. Diabetes Care 2017;40:832–8. http://dx.doi.org/10.2337/dc16-
1769.
[50] Bajaj HS, Venn K, Ye C, Patrick A, Kalra S, Khandwala H, et al. Lowest glucose
variability and hypoglycemia are observed with the combination of a GLP-1
receptor agonist and basal insulin (VARIATION study). Diabetes Care
2017;40:194–200. http://dx.doi.org/10.2337/dc16-1582.
[51] Franc S, Daoudi A, Pochat A, Petit M-H, Randazzo C, Petit C, et al. Insulin-based
strategies to prevent hypoglycaemia during and after exercise in adult
patients with type 1 diabetes on pump therapy: the DIABRASPORT random-
ized study. Diabetes Obes Metab 2015;17:1150–7. http://dx.doi.org/10.1111/
dom.12552.
[52] Wolpert HA. The nuts and bolts of achieving end points with real-time
continuous glucose monitoring. Diabetes Care 2008;31(Suppl. 2):S146–9.
http://dx.doi.org/10.2337/dc08-s238.
[53] Tubiana-Rufi N, Sierra P. La mesure continue du glucose chez l’enfant et
l’adolescent. Med Mal Metabol 2016;10:636–43.
[54] Wong JC, Foster NC, Maahs DM, Raghinaru D, Bergenstal RM, Ahmann AJ, et al.
Real-time continuous glucose monitoring among participants in the T1D
Exchange clinic registry. Diabetes Care 2014;37:2702–9. http://dx.doi.org/
10.2337/dc14-0303.
[55] Rewers MJ, Pillay K, de Beaufort C, Craig ME, Hanas R, Acerini CL, et al. ISPAD
clinical practice consensus guidelines 2014. Assessment and monitoring of
glycemic control in children and adolescents with diabetes. Pediatr Diabetes
2014;15(Suppl. 20):102–14. http://dx.doi.org/10.1111/pedi.12190.
[56] Phillip M, Danne T, Shalitin S, Buckingham B, Laffel L, Tamborlane W, et al. Use of
continuous glucose monitoring in children and adolescents (*). Pediatr Diabetes
2012;13:215–28. http://dx.doi.org/10.1111/j.1399-5448.2011.00849.x.
[57] Tubiana-Rufi N. Actualités des pompes et capteurs en pédiatrie. Med Mal
Metabol 2015;9.
[58] Telo GH, Volkening LK, Butler DA, Laffel LM. Salient characteristics of youth
with type 1 diabetes initiating continuous glucose monitoring. Diabetes
Technol Ther 2015;17:373–8. http://dx.doi.org/10.1089/dia.2014.0290.