diabetes research and clinical practice 1 4 7 (2 0 1 9) 1 5 7–16 5

Contents available at ScienceDirect

Diabetes Research
and Clinical Practice
journal homepage: www.elsevier.com/locat e/dia bre s

Similar glycaemic control with less nocturnal

hypoglycaemia in a 38-week trial comparing the
IDegAsp co-formulation with insulin glargine U100
and insulin aspart in basal insulin-treated subjects
with type 2 diabetes mellitus

A. Philis-Tsimikas a,*, K. Astamirova b, Y. Gupta c, A. Haggag d, D. Roula e, B.A. Bak f,

E.G. Fita f, A.M. Nielsen f, T. Demir g
a
Scripps Whittier Diabetes Institute, San Diego, CA, USA
b
Saint-Petersburg Territorial Diabetic Centre, Saint-Petersburg, Russia
c
All India Institute of Medical Sciences, New Delhi, India
d
Anaheim Clinical Trials, Anaheim, CA, USA
e
Salah Boubnider University, Constantine, Algeria
f
Novo Nordisk A/S, Søborg, Denmark
g _
Dokuz Eylül Üniversity, Izmir, Turkey

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

Article history: Aims: To confirm non-inferiority of insulin degludec/insulin aspart (IDegAsp) once-daily
Received 6 July 2018 (OD) versus insulin glargine (IGlar) U100 OD + insulin aspart (IAsp) OD for HbA1c after
Received in revised form 26 weeks, and compare efficacy and safety between groups at W26 + W38.
11 October 2018 Methods: A 38-week, randomised, open-label, treat-to-target (HbA1c < 7.0%) trial in adults
Accepted 29 October 2018 with type 2 diabetes mellitus (on basal insulin ± oral antidiabetic drugs; HbA1c 7.0–10.0%).
Available online 16 November 2018 Randomisation (1:1): IDegAsp or IGlar U100 + IAsp. Intensification to IDegAsp twice daily
(BID) was permitted at W26 + W32, or with additional IAsp injections at W26 (maximum
IAsp BID) or W32 (maximum IAsp three-times daily).
Keywords:
Results: For W0–W26, mean percentage-change (standard deviation) HbA1c was: IDegAsp,
IDegAsp
1.1 (0.9); IGlar U100 + IAsp, 1.1 (0.8); estimated treatment difference: 0.07% (95% confi-
Glargine
dence interval [CI]: 0.06; 0.21) confirmed non-inferiority. At W26 and W38, target HbA1c
Aspart
achievement, and mean fasting and postprandial glucose were similar across groups. At
Diabetes
W38, more subjects achieved target HbA1c without hypoglycaemia with IDegAsp (22.5%)
Co-formulation
than with IGlar U100 + IAsp (21.1%), with significantly fewer nocturnal episodes (W0–
Hypoglycaemia
W38, estimated rate ratio: 0.61 [95% CI: 0.40; 0.93]). Safety profiles were similar across treat-
ment groups throughout.
Conclusions: IDegAsp OD/BID are effective treatment intensification options versus multiple
injection basal–bolus therapies, achieving similar glycaemic control, with significantly less
nocturnal hypoglycaemia.
Ó 2018 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-
NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

* Corresponding author at: Scripps Whittier Diabetes Institute, 10140 Campus Point Dr, San Diego, CA 92121, USA.
E-mail address: [email protected] (A. Philis-Tsimikas).
https://doi.org/10.1016/j.diabres.2018.10.024
0168-8227/Ó 2018 The Authors. Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
158 diabetes research and clinical practice 1 4 7 ( 2 0 1 9 ) 1 5 7 –1 6 5

1. Introduction tors); HbA1c 7.0–10.0% (53–86 mmol/mol, both inclusive), and

body mass index 45 kg/m2.
Type 2 diabetes mellitus, a progressive disorder, is charac- Key exclusion criteria: treatment with any diabetes/obe-
terised by insulin resistance and impaired insulin secretion, sity medication not stated in the inclusion criteria; acute
resulting in chronic hyperglycaemia that worsens with dis- decompensation of glycaemic control requiring immediate
ease progression [1]. Several landmark studies demonstrate intensification of treatment to prevent severe metabolic dys-
the importance of maintaining tight glycaemic control, to regulation within 90 days prior to screening; myocardial
reduce the risk of long-term diabetes-related complications infarction, stroke or hospitalisation for unstable angina or
(e.g. UKPDS [2]; DCCT/EDIC [3]). transient ischaemic attack within 180 days prior to screening;
Current clinical guidelines from the American Diabetes New York Heart Association Class IV; planned coronary, caro-
Association (ADA) and European Association for the Study tid or peripheral artery revascularisation; and anticipated ini-
of Diabetes (EASD) follow a stepwise approach to treatment tiation or change in concomitant medications (for >14
intensification, combining lifestyle changes with pharmaco- consecutive days) known to affect weight/glucose
logical interventions [4,5]. If patients do not achieve target metabolism.
HbA1c with basal insulin, further treatment intensification is
recommended. Treatment intensification can involve addi- 3. Materials and methods
tional injections of bolus insulin for more tailored dosing,
addition of a glucagon-like peptide-1 receptor agonist (GLP-1 3.1. Trial design
RA), or use of combined basal and bolus products (which
may require fewer injections, and thus be more convenient A 38-week, randomised, open-label, multinational, treat-to-
for the patient). target trial (www.clinicaltrials.gov identifier: NCT02906917)
Insulin degludec/insulin aspart (IDegAsp) is a soluble co- in adults with type 2 diabetes mellitus, treated with basal
formulation of 70% insulin degludec (IDeg; providing basal insulin ± OADs and in need of treatment intensification.
coverage) and 30% insulin aspart (IAsp; providing post- The trial was conducted in seven countries, that each con-
prandial coverage), approved for the treatment of diabetes tributed the following number of randomised, exposed
mellitus in adults in many countries [6,7]. To date, no head- patients: Algeria: n = 46, Czech Republic: n = 60, India: n = 51,
to-head trials compare the effect of IDegAsp once-daily (OD) Russia: n = 110, Serbia: n = 46, Turkey: n = 65 and the United
therapy with a multiple injection basal–bolus regimen States: n = 150. An overview of the trial design is provided in
(involving separate and potentially escalating numbers of Fig. S1.
bolus insulin injections, together with a basal insulin).
The primary objective of this randomised, treat-to-target, 3.2. Randomisation, stratification and treatment
parallel-group, open-label trial was to confirm the effect of intensification
IDegAsp OD versus insulin glargine (IGlar) U100 OD + IAsp
OD on glycaemic control (HbA1c) after 26 weeks of treatment. All patients were randomised (1:1) from their previous basal
Additional objectives included: comparison of other efficacy insulin regimen to either: IDegAsp (Weeks 0–26: OD only;
and safety parameters after 26 weeks (IDegAsp OD versus Weeks 27–38: OD/BID), or IGlar U100 + IAsp (Weeks 0–26: IGlar
IGlar U100 OD + IAsp OD), and the comparison of other effi- U100 OD + IAsp OD only; Weeks 27–38: IGlar U100 OD + IAsp
cacy and safety parameters after 38 weeks (IDegAsp OD or OD/BID/TID). Patients were stratified equally into the two
twice-daily [BID] versus IGlar U100 OD + IAsp OD, BID, or treatment groups, based on pre-trial basal insulin regimens
three-times-daily [TID]). (OD versus BID/TID). Patients could be intensified (at physi-
cian discretion, based on the needs of the individual patient)
following intensification visits at Weeks 26 and 32, if HbA1c
2. Subjects was not on target in the previous week (target, <7%).
Titration data (e.g., self-measured blood glucose [SMBG];
2.1. Subject disposition and patient flow prescribed insulin dose; actual insulin units and date and
time given) were to be entered into the eCRF within 24 h of
From a total of 734 subjects who were screened, 532 fulfilled a site visit or home contact on weekdays for review by Novo
eligibility criteria and were randomised (IDegAsp, n = 267; Nordisk to reduce the time a subject may receive suboptimal
IGlar U100 + IAsp, n = 265) (Table S1). treatment.

2.2. Inclusion and exclusion criteria 3.3. Switching from previous treatment after
randomisation
Key inclusion criteria: male or female, aged 18 years; diag-
nosed with type 2 diabetes mellitus; treated for 90 days prior Patients on a prior OD regimen were shifted unit-to-unit,
to screening with any basal insulin either alone or with a whereas patients on a BID regimen received either IDegAsp
stable dose of oral antidiabetic drugs (OADs) (individual, or (unit-to-unit, previous dose), or IGlar U100 + IAsp (with a
any combination of: biguanides, sulphonylureas, glinides, 20% reduction of the total daily basal [IGlar U100] dose, and
dipeptidyl peptidase-4 [DPP-4] inhibitors, alpha-glucosidase IAsp initiated at 4 U with the largest meal). Sulphonylureas
inhibitors, or sodium/glucose cotransporter-2 [SGLT2] inhibi- and glinides were discontinued at randomisation; other OADs
 diabetes research and clinical practice 1 4 7 ( 2 0 1 9 ) 1 5 7 –1 6 5 159

continued unchanged. Diet and exercise counselling was pro- measurement at screening or randomisation, one or more
vided throughout the study for all subjects (as per standard of HbA1c measurement(s) after 12 weeks of exposure and a mini-
care at all investigational sites). mum of 12 weeks of exposure on-treatment; and the completer
analysis set (CAS), which included all randomised subjects
3.4. Details of treatment regimens who remained in the trial and on-treatment at Week 26.

IDegAsp OD/BID was administered with the largest meal(s) 3.7.2. Statistical analyses
each day. Investigators were trained during investigator meet- The primary estimand (‘‘estimand”, the estimated treatment
ings (through inspection of example meal plates) regarding effect on a given parameter) followed the ITT principle, and
how to best train patients to identify their ‘largest meal’ used the FAS to assess effectiveness in all randomised sub-
(nominally the meal containing the greatest amount of bread, jects (whether adherent to treatment or not), to determine
rice, potatoes, pasta or sugar) and encouraged to discuss the the non-inferiority of IDegAsp OD versus IGlar U100 OD + IAsp
identification process with patients during scheduled visits/- OD in terms of HbA1c reduction at Week 26, with a
phone contacts. After 26 weeks, if IDegAsp was intensified pre-specified non-inferiority margin of 0.4%. Subjects who
to be administered BID, one meal for dosing was to be evening discontinued treatment prior to Week 26 were included in
dinner; the other was either lunch or breakfast (depending on the analysis and missing data were imputed using multiple
which was identified as largest). Intensification from IDegAsp imputation (MI), with a separate imputation for each treat-
OD to BID was performed by splitting the total daily dose of ment group, based upon on- and off-treatment data. In the
IDegAsp OD into two doses (doses could be uneven, depend- IDegAsp group, all imputed values and off-treatment values
ing on the size of each meal). IGlar U100 was administered at Week 26 were penalised by 0.4%. Data were analysed using
OD in accordance with the local label throughout the trial. analysis of covariance (ANCOVA), with treatment, region, sex,
IAsp OD/BID/TID was administered with the largest meal(s), previous insulin treatment regimen, and previous OAD treat-
identified in the same manner as with IDegAsp. Rescue ther- ment as categorical fixed effects, and baseline response and
apy was permitted by the trial protocol, if patients exceeded age as covariates. Results were pooled using Rubin’s rule.
pre-specified blood glucose thresholds. The secondary estimand, in order to assess treatment
effects on all randomised subjects, applied a mixed model
3.5. Trial products for repeated measurements (MMRM) to the FAS. The MMRM
assumed subjects remained on-treatment throughout. Fur-
The following products were used throughout: IDegAsp 100 thermore, several sensitivity analyses were performed and
units (U)/mL (RyzodegÒ, Novo Nordisk A/S), delivered via a are displayed with a brief description of each analysis in
3 mL prefilled FlexTouchÒ pen (Novo Nordisk A/S); insulin Fig. S2. The sensitivity analyses included analyses of the PP
glargine 100 U/mL (LantusÒ, Sanofi S.A.), delivered via a and CAS analysis sets (accounting for missing data using MI
3 mL pre-filled SoloStarÒ pen (Sanofi S.A.); insulin aspart 100 or the MMRM), and an uncorrected estimate of the difference
U/mL (NovoRapidÒ/NovoLogÒ, Novo Nordisk A/S), delivered between the two treatments (the ‘‘sensitivity: MI, no non-
via a 3 mL prefilled FlexPenÒ pen (Novo Nordisk A/S). inferiority-penalty, FAS” analysis) performed as per the pri-
mary analysis, but without a 0.4% penalty for subjects in
3.6. Dose titration the IDegAsp arm.
The statistical analyses of total insulin dose, body weight,
Patients self-titrated to the following SMBG targets: IDegAsp fasting plasma glucose (FPG), SMBG and postprandial glucose
and IGlar U100, 4–5 mmol/L; IAsp, 4–6 mmol/L. The algo- increments were performed as per the primary analysis,
rithms used to determine dose adjustments (including details without applying a 0.4% penalty to the IDegAsp arm.
of the meals used for titration) are displayed in Table S2A Responder analyses (to determine achievement of HbA1c
(IDegAsp and IGlar U100) and Table S2B (IAsp). targets) were based on the FAS, with missing data imputed
using MI, with no penalty applied to the IDegAsp treatment
3.7. Statistical operations group. Subjects with HbA1c < 7% were considered ‘‘respon-
ders”; those with HbA1c  7% or with <16 weeks of treatment,
3.7.1. Data sets ‘‘non-responders”. Responder analyses were based on a logis-
The full analysis set (FAS) was used for the evaluation of effi- tic regression model which included treatment, sex, previous
cacy endpoints. The FAS contained all randomised subjects. insulin treatment regimen, previous OAD treatment, and
Statistical evaluation of the FAS followed the intention to region as fixed factors, and age and baseline HbA1c as
treat (ITT) principle, and subjects contributed ‘‘as covariates.
randomised”. Comparisons of hypoglycaemia between treatment groups
Safety outcomes were evaluated using the safety analysis were made using negative binomial regression, and included
set (SAS). The SAS contained all subjects who received one treatment, previous insulin treatment regimen, previous OAD
or more dose(s) of trial drug. Subjects in the SAS contributed treatment, sex, and region as factors and age as a covariate.
to the evaluation ‘‘as treated”.
Sensitivity analyses were conducted using the FAS and two 3.8. Hypoglycaemia classification
additional data sets: the per protocol analysis set (PP), which
included individuals from the FAS who did not violate any The evaluation of hypoglycaemia included severe (ADA 2013
inclusion criteria or fulfil any exclusion criteria, had an HbA1c definition: an episode requiring assistance of another person
160 diabetes research and clinical practice 1 4 7 ( 2 0 1 9 ) 1 5 7 –1 6 5

to actively administer carbohydrate, glucagon, or take other inhibitors (20% of each group). Few patients were in receipt
corrective actions; plasma glucose concentrations may not of SGLT2 inhibitors or alpha-glucosidase inhibitors (4–11% of
be available, however, neurological recovery following return each group), and <4% of patients in each group were taking
of plasma glucose to normal to be considered sufficient evi- combination OADs (i.e. >1 OAD combined in a single oral
dence of a severe episode [8]), or symptomatic blood medication). A total of four patients, randomised in error,
glucose-confirmed (<3.1 mmol/L) episodes. Nocturnal epi- were in receipt of OADs not permitted by the protocol (in each
sodes occurred 00.01–05.59 (both inclusive). group, one subject was on a GLP-1 RA and one was on a
thiazolidinedione).
4. Results
4.3. Description of treatment regimens throughout the
4.1. Subject disposition trial

The following number of subjects contributed to each data Fig. S3 provides an overview of the percentage of patients in
set, respectively: the FAS: IDegAsp, n = 267; U100 + IAsp, each treatment group assigned to OD/BID/TID regimens
n = 265; the SAS: IDegAsp, n = 265; U100 + IAsp, n = 263; the throughout Weeks 1–26, 27–31 and 32–38. Details of the num-
PP: IDegAsp, n = 250; U100 + IAsp, n = 248; and the CAS: ber of patients in each group who met the eligibility criteria
n = 243 from both treatment groups. for treatment intensification and who were/were not subse-
A total of 243 (91.0%) subjects in the IDegAsp group quently intensified at each assessment visit are in Table S5.
remained on-treatment at Week 26, and 235 (88.0%) com- The majority (>90%) of patients in both groups who achieved
pleted the trial on-treatment at Week 38. In the IGlar U100 target HbA1c of <7% were not intensified at either Week 26 or
+ IAsp group a total of 243 (91.7%) remained on-treatment at 32. For eligible patients not meeting the HbA1c target at Week
Week 26, and 236 (89.1%) completed the trial on-treatment 26, a high percentage of both groups had treatment intensi-
at Week 38. No patient required rescue therapy. More infor- fied (approximately 97% of each treatment group). At Week
mation on the reasons for treatment discontinuation and 32, approximately two-thirds of patients in both treatment
withdrawals is provided in Table S1. groups not reaching target in the previous week and eligible
for further intensification were intensified.
4.2. Baseline demographics and patient characteristics An overview of the dosing meals for each treatment group at
key time points (i.e. before and after treatment intensification
The baseline demographics of randomised subjects (Table 1) was permitted) is provided in Table S6. Prior to intensification,
were balanced between the treatment groups. The mean the majority of patients in each treatment group received OD
baseline HbA1c values were: IDegAsp, 8.2% (66 mmol/mol) dosing at dinner (49–55%), followed by lunch (30–38%), and
and IGlar U100 + IAsp, 8.1% (65 mmol/mol). Prior to randomi- then breakfast (13–15%). After intensification was permitted,
sation, the majority (>80%) of both treatment groups were on similar percentages of the IDegAsp group received BID doses at
an OD rather than a BID basal regimen and none were on a breakfast + dinner (27–29%) and lunch + dinner (25–28%),
TID basal regimen (Table S3). Most patients were being pre- whereas the IGlar U100 + IAsp group tended not to receive IAsp
scribed OADs (percentage in receipt of OADs at screening: doses at breakfast (most received IAsp doses at either lunch,
IDegAsp, 96.3%; IGlar U100 + IAsp, 92.5%), and the percentages dinner, lunch + dinner, or with all meals [Table S6]).
in receipt of one, two, or more than two OADs were well bal-
anced between groups (Table S3). 4.4. Number of daily injections
Table S4 provides information on the OAD regimens at
screening. The majority of patients were in receipt of bigua- At Week 26, the number of daily injections (as per protocol)
nides (>80% of each group). Sulphonylureas were also com- was: IDegAsp, one; IGlar U100 + IAsp, two. At Week 38, the
monly prescribed (>40% of each group), as were DPP-4 mean (standard deviation [SD]) number of daily injections for

Table 1 – Baseline demographic data.

IDegAsp, N = 267 IGlar U100 + IAsp, N = 265

Age (years) 58.2 ± 8.9 59.2 ± 9.1

Female (%) 53.2 48.3
Weight (kg) 88.6 ± 18.5 88.5 ± 17.5
BMI (kg/m2) 31.7 ± 5.5 31.7 ± 5.1
Diabetes duration (years) 12.9 ± 6.9 13.0 ± 6.5
HbA1c (%) [mmol/mol] 8.2 ± 0.8 [66 ± 8] 8.1 ± 0.7 [65 ± 8]
FPG (mmol/L) [mg/dL] 9.0 ± 2.7 [162 ± 48] 8.8 ± 2.7 [158 ± 48]
Data are mean ± standard deviation, or percentages.
BMI, body mass index; FPG, fasting plasma glucose; IDegAsp, insulin degludec/insulin aspart fixed-ratio combination once- or twice-daily; IGlar
U100 + IAsp, insulin glargine 100 units/mL once-daily + insulin aspart once-, twice- or three-times daily.
 diabetes research and clinical practice 1 4 7 ( 2 0 1 9 ) 1 5 7 –1 6 5 161

both groups had increased, to 1.62 ± 0.49 for the IDegAsp, and breakfast SMBG values were no longer statistically different
to 2.85 ± 0.87 for the IGlar U100 + IAsp groups, respectively. (ETD, 0.00 [95% CI: 0.22; 0.22]). For both treatment groups,
the shapes of the mean 24-h 9-point SMBG profiles (from
4.5. Efficacy results pre-breakfast, through to breakfast the following day) were
comparable, with a shift towards lower values as the trial pro-
Both treatment groups had similar HbA1c values throughout gressed (Fig. S4). Mean postprandial SMBG increments were
(Fig. 1). Across Weeks 0–26, similar changes in mean (SD) reduced in both groups after 26 and 38 weeks, with no statis-
HbA1c were observed in both groups: IDegAsp OD: 1.1% tical difference between groups found at either time point
(0.9); IGlar U100 OD + IAsp OD: 1.1% (0.8). The primary esti- (Table 2).
mand confirmed the non-inferiority of IDegAsp for change
in HbA1c: estimated treatment difference (ETD) 0.07% (95% 4.6. Insulin dose
confidence interval [CI]: 0.06; 0.21). The secondary estimand
(ETD 0.01% [95% CI: 0.12; 0.14]) and several sensitivity anal- The mean total insulin doses were similar at baseline
yses confirmed the robustness of the results in terms of gly- (IDegAsp, 35.2 U; IGlar U100 + IAsp, 34.6 U). Following
caemic control for 0–26 weeks (Fig. S2). treatment initiation, the total insulin doses diverged, with a
During Weeks 26–38, further treatment intensification was significantly lower dose associated with IDegAsp OD versus
permitted, and the HbA1c values of both groups continued to IGlar U100 OD + IAsp OD at Week 26 (70.9 U versus 79.4 U,
decrease (Fig. 1). At Week 38, there was no significant differ- respectively [a 10.7% lower dose]; odds ratio (OR) 0.88 U [95%
ence between the treatment groups in the primary and sec- CI: 0.81; 0.95]). By the end of the trial, the dose associated with
ondary estimands for HbA1c (Fig. S2). IDegAsp was significantly lower (6.6%) than that associated
Improvements in glycaemic control in both groups were with IGlar U100 + IAsp (83.4 U versus 89.3 U, respectively; OR
reflected by decreases in FPG, with similar profiles observed 0.91 U [95% CI: 0.83; 0.99]).
throughout Weeks 0–38 in both groups. At Week 26, the mean Differences in total insulin dose may be attributed to the
(SD) reduction of FPG was similar in both groups: IDegAsp OD: basal, rather than the bolus effect. At Week 38, the mean total
2.3 (2.9) mmol/L; IGlar U100 OD + IAsp OD: 2.3 (3.3) mmol/L. basal doses were: IDegAsp: 58.4 U; IGlar U100 + IAsp: 65.3 U,
At Week 38, IDegAsp was associated with a numerically and mean total bolus doses were: IDegAsp: 25 U; IGlar U100
greater mean (SD) reduction in FPG ( 2.7 [3.0]), versus 2.3 + IAsp: 24 U.
[3.1] mmol/L with IGlar U100 + IAsp); however, the ETD was
not statistically significant (Table 2). 4.7. Safety results
In Week 1, the mean pre-breakfast SMBG level (used for
titration) was higher with IDegAsp OD (8.8 mmol/L) than with 4.7.1. Hypoglycaemia
IGlar U100 OD + IAsp OD (8.3 mmol/L). At Week 26, the pre- No clustering of overall, or nocturnal, confirmed symptomatic
breakfast SMBG values were statistically higher with IDegAsp hypoglycaemic episodes were observed in the first few weeks
(ETD, 0.28 [95% CI: 0.05; 0.51]). By Week 38, however, the pre- following randomisation: Fig. 2A and B.

Fig. 1 – Change in HbA1c over the treatment period. Dashed line shows start of treatment intensification (permitted Weeks 26–
38). Error bars show standard error of the mean. Primary effectiveness estimand (full analysis set; missing data imputed
using multiple imputation; pre-specified non-inferiority margin at Week 26: 0.4%). CI, confidence interval; ETD, estimated
treatment difference; IDegAsp, insulin degludec/insulin aspart fixed-ratio combination once- or twice-daily; IGlar U100+IAsp,
insulin glargine 100 units/mL once-daily + insulin aspart once-, twice- or three-times daily.
 162
Table 2 – Key endpoints.
Treatment initiation (Weeks 0–26) Baseline–end of treatment (Weeks 0–38)
†
IDegAsp (N = 267) IGlar U100 + IAsp (N = 265) ETD, OR or RR (95% CI) IDegAsp (N = 267) IGlar U100 + IAsp (N = 265) ETD, OR or RR† (95% CI)

Change from 1.1 (0.9) 1.1 (0.8) ETD 1.3 (0.8) 1.2 (0.8) ETD
baseline HbA1c 0.07 0.09
(%), mean (SD) ( 0.06; 0.21) ( 0.04; 0.22)

Change from 12 12 ETD 14 14 ETD

baseline HbA1c (9) (9) 0.82 (9) (9) 0.95
(mmol/L), mean ( 0.64; 2.28) ( 0.48; 2.38)

diabetes research and clinical practice

(SD)

Percentage of 44.9 45.3 OR 52.1 52.8 OR

HbA1c responders 1.07 (0.74; 1.54) 0.95
(<7%) (0.66; 1.38)

Overall 258.5 296.1 RR 287.1 343.3 RR

hypoglycaemia,à 0.90 (0.67; 1.22) 0.86 (0.65; 1.14)
mean rate
(events/100 PYE)

Nocturnal 47.9 92.9 RR 60.4 101.4 RR

hypoglycaemia,à 0.55 (0.34; 0.90) 0.61 (0.40; 0.93)
mean rate
(events/100 PYE)

Total daily insulin 70.9 (41.5) 79.4 (37.7) OR 83.4 (51.3) 89.3 (43.1) OR
dose (units), 0.88 0.91
mean (SD) (0.81; 0.95) (0.83; 0.99)

1 4 7 ( 2 0 1 9 ) 1 5 7 –1 6 5
Change from 2.3 (2.9) 2.3 (3.3) ETD 2.7 (3.0) 2.3 (3.1) ETD
baseline FPG 0.04 0.24
(mmol/L), mean ( 0.34; 0.42) ( 0.60; 0.13)
(SD)

Change from 0.6 (2.6) 0.5 (2.3) ETD 1.0 (2.2) 1.1 (2.5) ETD
baseline SMBG 0.10 0.30
postprandial ( 0.23; 0.43) ( 0.01; 0.62)
increment (mmol/
L), mean (SD)

Total number of 1 (0.00) 2 (0.00) Not tested 1.62 (0.49) 2.85 (0.87) Not tested
injections/day,
mean (SD)
Data are observed values except ETD, OR and RR (estimated values). †ETD = IDegAsp – IGlar U100 + IAsp; RR = IDegAsp/IGlar U100 + IAsp. àHypoglycaemia included severe (American Diabetes
Association 2013 definition: an episode requiring assistance of another person to actively administer carbohydrate, glucagon, or take other corrective actions) or blood glucose-confirmed
(<3.1 mmol/L) symptomatic episodes. Nocturnal episodes occurred 0.01–05.59 (both inclusive). CI, confidence interval; ETD, estimated treatment difference; FPG, fasting plasma glucose; IDegAsp,
insulin degludec/insulin aspart fixed-ratio combination once- or twice-daily; IGlar U100 + IAsp, insulin glargine 100 units/mL once-daily + insulin aspart once-, twice- or three-times daily; OR, odds
ratio; PYE, patient–years of exposure; RR, estimated rate ratio; SD, standard deviation; SMBG, self-measured blood glucose.
 diabetes research and clinical practice 1 4 7 ( 2 0 1 9 ) 1 5 7 –1 6 5 163

During treatment initiation (Week 0–26) there were numer- Rates of severe hypoglycaemic episodes were low through-
ically fewer overall confirmed symptomatic episodes per sub- out, with seven episodes reported in the IDegAsp group and
ject associated with IDegAsp OD compared with IGlar U100 nine in the IGlar U100 + IAsp group. Additional information
OD + IAsp OD; however, the difference was not statistically on hypoglycaemia throughout the trial is provided in Table S7.
significant (Fig. 2A). In the same period, there were signifi- At Weeks 26 and 38, the percentage of patients with
cantly fewer nocturnal confirmed symptomatic episodes per HbA1c on-target (<7.0%) who did not experience a con-
subject associated with IDegAsp (estimated rate ratio [RR] firmed symptomatic hypoglycaemic episode was 27.3%
0.55 [95% CI: 0.34; 0.90]; a 45% rate-reduction versus IGlar and 22.5%, respectively, for IDegAsp compared with 23.0%
U100 + IAsp [Fig. 2B]). and 21.1%, respectively, for IGlar U100 + IAsp. Prior to the
Similarly, when the entire treatment period (Weeks 0–38) treatment intensification period at Week 26, when noctur-
was considered, there were numerically fewer overall con- nal episodes were considered, the percentage of patients
firmed symptomatic hypoglycaemic episodes per subject achieving target HbA1c without nocturnal hypoglycaemia
associated with IDegAsp compared with IGlar U100 + IAsp, was 40.8% for IDegAsp and 34.0% for IGlar U100 + IAsp
but the difference between treatment groups was not statisti- (OR: 1.49 [95% CI: 1.03; 2.17]). At Week 38, the significant
cally significant (Fig. 2A). In the same period there were signif- difference between the groups was no longer observed with
icantly fewer nocturnal confirmed symptomatic episodes per the percentage of patients achieving target HbA1c without
subject associated with IDegAsp compared with IGlar U100 nocturnal hypoglycaemia being similar in both treatment
+ IAsp (RR 0.61 [95% CI: 0.40; 0.93]; a 39% rate-reduction versus groups: IDegAsp: 38.6%; IGlar U100 + IAsp: 35.5% (OR: 1.18
IGlar U100 + IAsp [Fig. 2B]). [95% CI: 0.81; 1.71]).

Fig. 2 – Rates of symptomatic blood glucose-confirmed hypoglycaemia. Panel A. all episodes; Panel B. nocturnal episodes
(00.01–05.59, both inclusive). Graphs: safety analysis set, comparisons (full analysis set) were made using negative binomial
regression. Dashed line shows start of treatment intensification (permitted Weeks 26–38). CI, confidence interval; IDegAsp,
insulin degludec/insulin aspart fixed-ratio combination once- or twice-daily; IGlar U100+IAsp, insulin glargine 100 units/mL
once-daily + insulin aspart once-, twice- or three-times daily; RR, estimated rate ratio (IDegAsp/IGlar U100 + IAsp).
164 diabetes research and clinical practice 1 4 7 ( 2 0 1 9 ) 1 5 7 –1 6 5

The lower rates of confirmed symptomatic nocturnal This trial is important, as it offers clinical guidance on
hypoglycaemia associated with IDegAsp (events per 100 how, when and why to use IDegAsp in patients with type
patient–years of exposure at Week 38: IDegAsp: 60.42 events; 2 diabetes mellitus not achieving target HbA1c with basal
IGlar U100 + IAsp: 101.38 events [Table S7]) translated into insulin. The IDegAsp treatment modality was compared
the following numbers needed to treat for 1 year with IDe- with an established multiple injection basal–bolus regimen
gAsp versus treatment with IGlar U100 + IAsp, in order to (considered to be more complex, due to requirements for
avoid one symptomatic nocturnal hypoglycaemic episode: more injections and use of two titration algorithms). IDe-
2.40 patients at Week 26; 2.52 patients at Week 38. gAsp was titrated according to a ‘‘precise sliding scale”
algorithm, using pre-breakfast SMBG goals. The IDegAsp
4.7.2. Adverse events algorithm was similar to the basal treatment algorithm,
There were similar rates of treatment-emergent adverse maintaining familiarity and simplicity for the patient,
events reported in both treatment groups. The majority were through a unit-to-unit transfer.
neither serious, nor judged to be related to trial products The results of this trial also offer specific guidance on the
(Table 3). There were two deaths reported in the IGlar U100 further intensification of IDegAsp OD, splitting the dose when
+ IAsp group, both were considered unrelated to the trial necessary, and demonstrating that IDegAsp is as effective for
products (one subject had pancreatic carcinoma metastatic with intensification of treatment as a multiple injection basal–
subsequent hepatic metastases, intra-abdominal lymph node bolus regimen, with a more desirable nocturnal hypoglycaemia
metastasis and cancer of the tail of the pancreas; the other, profile. Furthermore, these data may be relevant to general
myocardial ischaemia as a result of ischaemic heart disease practitioners treating patients with type 2 diabetes mellitus
[however, possibility of a severe hypoglycaemic episode could in need of intensification. IDegAsp represents a simple,
not be excluded as an alternative cause of death by the adaptable treatment option requiring a single injection,
investigator]). which may better suit patient lifestyle compared with more
complex alternatives, with a lower daily injection frequency
4.7.3. Body weight (compared with basal–bolus regimens) likely to be perceived
Similar mean baseline body weights were observed in both by patients as less burdensome [9].
treatment groups (IDegAsp: 88.6 kg; IGlar U100 + IAsp: Nocturnal hypoglycaemia is a widely recognised and
88.5 kg, Table 1), and the increases in body weight observed feared complication of diabetes mellitus, representing a bar-
after 38 weeks were similar for both treatment groups: IDe- rier to treatment intensification for some patients, with a
gAsp: +2.5 kg; IGlar U100 + IAsp: +2.4 kg. No statistical differ- detrimental impact on quality of life and/or serious clinical
ences in body weight were found at Week 26 (ETD 0.43 [95% consequences [10–12]. It is therefore of importance that noc-
CI: 0.13; 0.99]), or Week 38 (ETD 0.16 [95% CI: 0.47; 0.79]). turnal hypoglycaemia was significantly reduced with IDe-
gAsp, compared with IGlar U100 + IAsp. Furthermore, it was
5. Discussion reassuring no safety concerns were associated with the
switch from pre-trial regimens, with no clustering of hypogly-
This trial confirmed a similar glycaemic effect for IDegAsp OD caemic episodes noted in the weeks immediately post-
versus IGlar U100 OD + IAsp OD over 26 weeks of treatment, randomisation in either group. It is likely that nocturnal
with fewer injections, significantly less nocturnal hypogly- hypoglycaemic events are attributable to the
caemia and a lower insulin dose associated with IDegAsp. pharmacokinetic/pharmacodynamic (PK/PD) properties of
These findings were consistent over the 38-week treatment the basal insulin component (IGlar U100) [13]. It has been pre-
period, where IDegAsp OD/BID provided equivalent glycaemic viously demonstrated that insulin degludec provides a very
control, with statistically significantly less nocturnal hypogly- flat glucose-lowering PK/PD profile with low variability [14],
caemia, fewer injections, and a lower insulin dose compared and that this translates into a risk reduction for nocturnal
with IGlar U100 OD + IAsp OD/BID/TID. hypoglycaemia versus IGlar [15].

Table 3 – Treatment-emergent adverse events (Week 0–38).

IDegAsp, N = 265 IGlar U100 + IAsp, N = 263
N (%) E R N (%) E R

Adverse event 175 (66.0) 614 328.30 168 (63.9) 525 281.61
Serious 18 (6.8) 24 12.83 20 (7.6) 26 13.95
Fatal* 0 – – – 2 (0.8) 2 1.07
Not recovered 68 (25.7) 129 68.97 71 (27.0) 124 66.51
Severe 9 (3.4) 14 7.49 11 (4.2) 18 9.66
Probably/possibly related to trial product 31 (11.7) 84 44.91 22 (8.4) 46 24.67
Safety analysis set. *Fatal events: myocardial ischaemia and pancreatic carcinoma metastatic. E, number of events; IDegAsp, insulin
degludec/insulin aspart fixed-ratio combination once- or twice-daily; IGlar U100 + IAsp, insulin glargine 100 units/mL once-daily + insulin
aspart once-, twice- or three-times daily; N, number of subjects with 1 event; R, rate per 100 patient–years of exposure; %, percentage of
subjects with 1 event.
 diabetes research and clinical practice 1 4 7 ( 2 0 1 9 ) 1 5 7 –1 6 5 165

The current trial had several strengths. The trial design [3] Diabetes Control and Complications Trial (DCCT)/
was close to routine clinical practice, with the treat-to- Epidemiology of Diabetes Interventions and Complications
target aspect allowing optimisation of glycaemic control for (EDIC) Study Research Group. Intensive diabetes treatment
and cardiovascular outcomes in type 1 diabetes: The DCCT/
patients on an individual basis, and selecting IGlar U100 as
EDIC study 30-year follow-up. Diabetes Care 2016;39:686–93.
a comparator enabled comparison with one of the most [4] Inzucchi SE, Bergenstal RM, Buse JB, Diamant M, Ferrannini E,
widely used basal insulins [12]. In addition, high completion Nauck M, et al. Management of hyperglycaemia in type 2
rates were recorded for both treatment groups. Limitations diabetes, 2015: a patient-centred approach. Update to a
included the open-label design as a potential source of bias position statement of the American Diabetes Association and
(however, it should be noted that blinding in such a trial the European Association for the Study of Diabetes.
Diabetologia 2015;58:429–42.
would not be feasible due to the number of injections differ-
[5] American Diabetes Association. 8. Pharmacologic
ing between the treatment groups).
approaches to glycemic treatment: standards of medical care
In summary, the trial confirms IDegAsp OD is as effective in diabetes-2018. Diabetes Care 2018;41:S73–S85.
as the IGlar U100 + IAsp multiple injection basal–bolus regi- [6] European Medicines Agency (EMA). Ryzodeg summary of
men in terms of glycaemic control, but with significantly less product characteristics; 2017. Available at: <http://www.ema.
nocturnal hypoglycaemia. Furthermore, stepwise intensifica- europa.eu/docs/en_GB/document_library/EPAR_-_Product_
tion from IDegAsp OD to BID seems feasible with respect to Information/human/002499/WC500139011.pdf> [Last
accessed: 21 June 2018].
both glycaemic control and safety.
[7] United States Food and Drug Administration (FDA). Ryzodeg
approval letter; 2015. Available at: <https://www.accessdata.
Acknowledgements fda.gov/drugsatfda_docs/appletter/2015/203313Orig1s000ltr.
pdf> [Last accessed: 21 June 2018].
Medical writing and editing assistance was provided by James [8] Seaquist ER, Anderson J, Childs B, Cryer P, Dagogo-Jack S, Fish
Currie and Grant Womack of Watermeadow Medical, an Ash- L, et al. Hypoglycemia and diabetes: a report of a workgroup
of the American Diabetes Association and the Endocrine
field Company, part of UDG Healthcare plc, funded by Novo
Society. Diabetes Care 2013;36:1384–95.
Nordisk A/S. This study was supported by Novo Nordisk.
[9] Polonsky WH, Fisher L, Hessler D, Bruhn D, Best JH. Patient
perspectives on once-weekly medications for diabetes.
Declarations of interest Diabetes Obes Metab 2011;13:144–9.
[10] Brod M, Christensen T, Bushnell DM. Impact of nocturnal
AP-T has been a researcher and advisor for AstraZeneca, Dex- hypoglycemic events on diabetes management, sleep quality,
and next-day function: results from a four-country survey. J
com, Lilly, Novo Nordisk and Sanofi, and an advisor for Merck.
Med Econ 2012;15:77–86.
DR has received personal fees and non-financial support from
[11] Edelman SV, Blose JS. The impact of nocturnal hypoglycemia
Hikma, Novo Nordisk and Sanofi. BAB, EGF and AMN are on clinical and cost-related issues in patients with type 1 and
employees of Novo Nordisk. KA, YG, AH and TD have no con- type 2 diabetes. Diabetes Educ 2014;40:269–79.
flicts of interest to declare. [12] Standl E, Owen DR. New long-acting basal insulins: does
benefit outweigh cost? Diabetes Care 2016;39(Suppl 2):S172–9.
[13] Goldman J, Kapitza C, Pettus J, Heise T. Understanding how
Appendix A. Supplementary material pharmacokinetic and pharmacodynamic differences of basal
analog insulins influence clinical practice. Curr Med Res Opin
Supplementary data to this article can be found online at 2017;33:1821–31.
https://doi.org/10.1016/j.diabres.2018.10.024. [14] Heise T, Hovelmann U, Nosek L, Hermanski L, Bottcher SG,
Haahr H. Comparison of the pharmacokinetic and
pharmacodynamic profiles of insulin degludec and insulin
glargine. Expert Opin Drug Metab Toxicol 2015;11:1193–201.
R E F E R E N C E S
[15] Ratner RE, Gough SC, Mathieu C, Del Prato S, Bode B,
Mersebach H, et al. Hypoglycaemia risk with insulin degludec
compared with insulin glargine in type 2 and type 1 diabetes:
[1] American Diabetes Association. 2. Classification and a pre-planned meta-analysis of phase 3 trials. Diabetes Obes
diagnosis of diabetes: standards of medical care in diabetes- Metab 2013;15:175–84.
2018. Diabetes Care 2018;41:S13–S27.
[2] Holman RR, Paul SK, Bethel MA, Matthews DR, Neil HA. 10-
year follow-up of intensive glucose control in type 2 diabetes.
N Engl J Med 2008;359:1577–89.