Article: Treatment

A study comparing insulin lispro mix 25 with glargine

plus lispro therapy in patients with Type 2 diabetes
who have inadequate glycaemic control on oral anti-
hyperglycaemic medication: results of the PARADIGM
study
K. Bowering
1
, V. A. Reed
2
, J. Felicio
3
, J. Landry
4
, L. Ji
5
and J. Oliveira
6
1
University of Alberta, Division of Endocrinology and Metabolism in the Department of Medicine, Edmonton, AB, Canada,
2
Eli Lilly and Company, Asia-Pacic
Medical Communications, Sydney, NSW, Australia,
3
Universidade Federal do Para , Endocrinology Division, Bele m, Brazil,
4
Eli Lilly Canada Inc., Toronto, ON,
Canada,
5
Peking University Peoples Hospital, Department of Endocrinology and Metabolism, Beijing, China and
6
Eli Lilly and Company, Indianapolis, IN, USA
Accepted May 30 2012
Abstract
Aims To test the hypothesis that initiation and intensication with 25% insulin lispro, 75% insulin lispro protamine
suspension (LM25), is non-inferior to initiation and intensication with glargine + insulin lispro therapy on change from
baseline in HbA
1c
.
Methods In this randomized, non-inferiority (margin of 0.4%), parallel, prospective, multi-country, 48-week, open-label
study, patients (n = 426) with Type 2 diabetes inadequately controlled with oral anti-hyperglycaemic medications were
assigned to either initiating therapy with one daily LM25 injection, progressing up to three daily injections (full analysis set
n = 211; per protocol set n = 177) or initiating therapy with one daily glargine injection and progressing up to three daily
insulin lispro injections (full analysis set n = 212; per protocol set n = 184).
Results LM25 therapy was found to be non-inferior to glargine + insulin lispro therapy by study end (upper limit of
95% CI < 0.4), with a least-squares mean difference (95% CI) in HbA
1c
(LM25 minus glargine + insulin lispro) of
)0.4 mmol mol (95% CI )2.7 to 1.9); )0.04% (95% CI )0.25 to 0.17). No statistically signicant differences between
treatment groups were found in the percentage of patients achieving HbA
1c
targets or postprandial blood glucose levels. The
increase in insulin dose, number of injections and weight change during the course of the study were similar in both groups.
Patients in both groups experienced similar hypoglycaemia rates and safety prole.
Conclusions For patients with Type 2 diabetes inadequately controlled with oral anti-hyperglycaemic medications,
glycaemic control when initiating and intensifying with LM25 therapy was found to be non-inferior to treatment with
glargine + insulin lispro therapy.
Diabet. Med. 29, e263e272 (2012)
Keywords clinical trial, insulin mixtures, Type 2 diabetes
Abbreviations LM25, insulin lispro mix (25% insulin lispro, 75% insulin lispro protamine suspension); LM50, insulin
lispro mix (50% insulin lispro, 50% insulin lispro protamine suspension)
Introduction
When patients with Type 2 diabetes fail to gain optimal gly-
caemic control with diet, exercise and oral anti-hyperglycaemic
medications, they may require insulin therapy, as long-term
glycaemic control reduces the risks of both microvascular and
macrovascular complications [1,2]. Possible options for these
patients initiating insulin therapy include basal insulin with
subsequent addition of lispro and premixed insulin regimens
[3,4]. Basalbolus insulin treatment may be more complicated
for patients than premixed insulin therapy, typically requiring
Correspondence to: Juliana Oliveira, Eli Lilly and Company, Lilly Corporate
Centre, Indianapolis, IN 46285, USA. E-mail: juliana_oliveira@lilly.com
(Clinical Trials Registry No; NCT 00548808)
DIABETICMedicine
DOI: 10.1111/j.1464-5491.2012.03722.x
2012 Eli Lilly and Company.
Diabetic Medicine 2012 Diabetes UK e263
more daily injections, different types of insulin and injection
devices, and more rigorous monitoring [5].
Relatively few head-to-head comparisons of the efcacy
and safety of premixed insulin therapy have been reported
[612]. Two head-to-head studies have been reported, com-
paring therapy with premixed 50% basal (insulin lispro
protamine suspension) and 50% prandial (insulin lispro)
insulin to basalbolus therapy. These studies, however, failed
to show non-inferiority of premixed insulin compared with
basalbolus therapy, the gold standard of insulin treat-
ment, although patients treated with premixed insulin did
signicantly improve their glycaemic control from baseline
[7,12].
An alternative premixed option for patients is to initiate
therapy with a premixed insulin with a low prandial:basal
insulin ratio. A single-arm study investigated initiation and
intensication of therapy with 70% insulin aspart protamine
suspension and 30% insulin aspart in patients with Type 2
diabetes failing on oral anti-hyperglycaemic medications (with
or without basal insulin), initiating with one daily injection
and progressing up to three daily injections, by adding one
injection every 16 weeks [13]. In this study, Garber and col-
leagues reported that 60% of patients achieved HbA
1c
of
48 mmol mol ( 6.5%) and 77% achieved HbA
1c
of
53 mmol mol ( 7.0%). However, this study had a number
of limitations, including the absence of a control or comparator
arm.
The purpose of this head-to-head, randomized study,
therefore, was to compare a low prandial ratio premixed
insulin [75% insulin lispro protamine suspension, 25% insulin
lispro (LM25)] with one daily glargine injection and pro-
gressing up to three additional daily insulin lispro injections
(glargine + lispro) as needed to meet glycaemic targets. The
primary objective of this study was to test, in patients with
Type 2 diabetes inadequately controlled with oral anti-
hyperglycaemic medications, whether initiation and intensi-
cation with LM25 therapy was non-inferior to initiation and
intensication with glargine plus lispro, in terms of glycaemic
control.
Research design and methods
This 48-week, randomized, open-label, active-controlled study
was conducted in accordance with the International Confer-
ence on Harmonisation Guidelines for Good Clinical Practice
and the Declaration of Helsinki [14]. Patients were enrolled in
Australia, Brazil, Canada, China, India, South Korea and
Mexico, from November 2007 to August 2009. Written
informed consent was provided by all patients.
Eligible patients were men and women age 3080 years with
Type 2 diabetes with inadequate glycaemic control [HbA
1c
53 mmol mol ( 7.0) and < 97 mmol mol (< 11.0%)] while
taking oral anti-hyperglycaemic medications (metformin plus
sulphonylurea and or pioglitazone) without insulin, for at least
90 days [15].
Patients were excluded if they were taking any glucose-
lowering agents, with the exception of metformin, pioglitazone
or sulphonylurea. Pioglitazone was discontinued 2 weeks prior
to randomization in countries where the concomitant use with
insulin was contraindicated. Patients with a BMI 35 kg m
2
,
functional capacity class III IV cardiac disease, impaired renal
function, active liver disease, or more than one episode of
severe hypoglycaemia within 24 weeks prior to study entry,
were also excluded.
Study medications and treatments
Eligible patients were randomly assigned to either the LM25
treatment group (initiating insulin therapy with one daily
LM25 injection and progressing up to three daily injections) or
the glargine + lispro treatment group (initiating with one daily
glargine injection and progressing up to three additional daily
insulin lispro injections) in a 1:1 ratio, and stratied by country,
HbA
1c
[ 69 mmol mol ( 8.5%) and > 69 mmol mol
(> 8.5%)] and continuing use of a sulphonylurea.
Patients started insulin therapy with one 10-U injection of
LM25 (administered within 15 min prior to the evening meal)
or with one 10-U injection of insulin glargine (administered at
bedtime), depending on treatment group assignment.
The insulin intensication ow chart is described in Fig. 1.
This titration process was followed by investigators during the
course of the study. Insulin doses, self-monitored blood glucose
values and any events associated with signs or symptoms of
hypoglycaemia were recorded in patient diaries.
The dose and number of injections were based on patients
glycaemic needs as reected by self-monitored blood glucose
measurements. Blood glucose concentration targets for fasting
and preprandial conditions were less than 6.1 mmol l
(110 mg dl) (Roche plasma, Roche Diagnostics, Indianapolis,
IN, USA). Patients could adjust the dose regimen at any time
during the study according to the dose titration algorithms
described in Table 1 (LM25 injections) and Table 2 (insulin
glargine and insulin lispro). Doses were adjusted based on pre-
meal values; for example, the dose at breakfast was based on
the pre-lunch blood glucose value. Patients were advised to
allow 34 days before making a subsequent dose change.
Throughout the study, the HbA
1c
level and the fasting blood
glucose levels were not blinded to the investigators; dose
adjustments were based on these values.
Investigators encouraged patients to establish a stable insulin
dose (at least 4 weeks) and to lower blood glucose values to
near target before adding an additional injection. Additional
injections were initiated by the investigator at one of the
weekly in-person visits up to and including week 8, or in a
biweekly telephone call from week 8 to week 48. Sulpho-
nylureas were discontinued before adding the second injection
of LM25 in the LM25 group, or the rst injection of insulin
lispro in the glargine + lispro group. Patients assigned to the
LM25 treatment group could add up to two additional
injections, and patients assigned to the glargine + lispro
DIABETICMedicine Lispro 25 vs. glargine plus lispro therapy in Type 2 diabetes K. Bowering et al.
2012 The Authors.
e264 Diabetic Medicine 2012 Diabetes UK
treatment group could add up to three injections of insulin
lispro.
Safety was monitored throughout the study. Events related to
hypoglycaemia were assessed as to incidence, rate and severity.
Hypoglycaemia was dened as any time a patient felt that
he she was experiencing a sign or symptom associated with
hypoglycaemia or had a blood glucose level of < 3.9 mmol l
(70 mg dl), even if it was not associated with signs, symptoms
or treatment. Severe hypoglycaemia was dened as an episode
with symptoms consistent with hypoglycaemia in which the
patient required the assistance of another person, and was
associated with either a blood glucose level of < 2.8 mmol l
(50 mg dl) or prompt recovery after oral carbohydrate, glu-
cagon or intravenous glucose. Nocturnal hypoglycaemia was
dened as any hypoglycaemic event that occurred between
bedtime and waking.
Outcome measures
The primary efcacy measure was change in HbA
1c
from
baseline to endpoint. Secondary outcome measures were: per-
centage of patients achieving HbA
1c
targets of 48 mmol mol
( 6.5%) and < 53 mmol mol (< 7.0%) at endpoint;
postprandial blood glucose and daily total insulin dose; 7-point
self-monitored blood glucose proles; and lipid and choles-
terol proles. Planned exploratory objectives were: patient
evaluation of quality of life, as assessed by the EuroQol
instrument (EQ-5D); and patient evaluation of disease-specic
FIGURE 1 Insulin intensication owchart. *Doses were adjusted based on pre-meal values; for example, the dose at breakfast was based on the pre-lunch
blood glucose value. Patient-made dose adjustments were assessed by the investigator at least weekly for the rst 8 weeks, then at least once every 2 weeks
for the remainder of the study. Investigators encouraged patients to establish a stable insulin dose (at least 4 weeks) and to lower blood glucose values to
near target before adding an additional injection. Additional injections were initiated by the investigator weekly up to and including week 8 and biweekly
to week 48. Sulphonylureas were discontinued before adding the second injection of LM25 in the LM25 group, or the rst injection of insulin lispro in the
glargine + lispro group. LM25, insulin lispro mix (25% insulin lispro, 75% insulin lispro protamine suspension).
DIABETICMedicine Original article
2012 Eli Lilly and Company.
Diabetic Medicine 2012 Diabetes UK e265
quality of life, as assessed by the Diabetes Health Prole (DHP-
18).
Statistical methods
The sample size was calculated based on the primary efcacy
measure of change in HbA
1c
from baseline to endpoint.
Assuming a standard deviation of 1.1% and dropout rate of
15%, 200 randomly assigned patients (160 completing the
study) in each treatment group would permit conrmation of
non-inferiority using an upper limit of 0.4% with a 95%
condence interval (CI) and 90% power.
Efcacy measures were analysed using the per protocol
analysis set, which included patients who completed
32 weeks of study treatment and had no major protocol
violations. Exclusions from analyses were based on the prede-
termined exclusionary criteria of the protocol. The intention-to-
treat set was applied as supportive analyses to all of the per
protocol analyses. The intention-to-treat set comprised patients
who had at least one dose of study drug.
Safety measures were analysed using the full analysis set,
which included all patients who received at least one dose of
study drug. The study endpoint was dened as the last patient
visit after completing 32 weeks of study treatment.
The primary analysis was conducted using an analysis of
covariance (ANCOVA) with baseline HbA
1c
as a covariate, and
treatment group, country and sulphonylurea use as factors.
Least-squares means with two-sided 95%condence interval of
the treatment difference were generated from the model using
type III sums of squares. If the upper limit of the 95% con-
dence interval was below0.4%(LM25 minus glargine + lispro),
then LM25 was non-inferior to glargine + lispro. Supportive
Table 1 Insulin dose titration algorithms for once-, twice- and thrice-daily LM25
Once-daily LM25* Twice-daily LM25 Thrice-daily LM25
Fasting blood
glucose mmol l
(mg dl)
Next day pre-evening
meal dose change
Pre-evening
meal fasting blood
glucose mmol l
(mg dl)
Next day pre-breakfast
pre-evening meal dose
change
Pre-evening meal
blood glucose mmol l
(mg dl)
Next day pre-lunch
dose change
< 2.8 (< 50) )4 units < 2.8 (< 50) )4 units < 2.8 (< 50) )3 units
2.84.4 (5079) )2 units 2.84.4 (5079) )2 units 2.84.4 (5079) )2 units
4.56.0 (80109) No change 4.56.0 (80109) No change 4.56.0 (80109) )1 unit
6.17.7 (110139) +2 units 6.17.7 (110139) +2 units 6.17.7 (110139) No change
7.811.0 (140199) +4 units 7.8 ( 140) +4 units 7.8 ( 140) +2 units
> 11.0 (> 199) +6 units
LM25, insulin lispro mix (25% insulin lispro, 75% insulin lispro protamine suspension).
*LM25 therapy was initiated with one 10-U injection administered within 15 min prior to the evening meal.
Additional injections of LM25 were added to the patients regimen when HbA
1c
and or blood glucose targets were not achieved with the
existing insulin injections.
A second LM25 injection, if required, was administered before breakfast, composed of half of the previous total pre-evening meal dose.
A third injection of 3-U of LM25, if required, was administered before lunch.
Patients taking a sulphonylurea discontinued prior to adding the second LM25 injection.
Table 2 Insulin dose titration algorithms for insulin glargine and once-, twice- and thrice-daily insulin lispro
Once-daily glargine* Up to thrice-daily insulin lisprox
Fasting blood
glucose mmol l
(mg dl)
Next day bedtime
dose change
Pre-meal bedtime
blood glucose mmol l (mg dl)
Next day
dose change
< 4.5 (< 80) )2 units < 2.8 (< 50) )4 units
4.56.0 (80109) No change 2.84.4 (5079) )2 units
6.16.6 (110120) +2 units 4.56.0 (80109) No change
6.77.7 (121140) +4 units 6.17.7 (110139) +2 units
7.88.8 (141160) +6 units 7.8 ( 140) +4 units
8.9 ( 161) +8 units
*Therapy was initiated with one 10-U injection of insulin glargine administered at bedtime.
Insulin lispro injections were added one at a time in order to target the higher blood glucose values rst.
Additional prandial injections of insulin lispro were added to the patients regimen as needed, of approximately 10% of the total daily units
of insulin glargine administered.
Patients taking a sulphonylurea discontinued prior to adding an insulin lispro injection.
DIABETICMedicine Lispro 25 vs. glargine plus lispro therapy in Type 2 diabetes K. Bowering et al.
2012 The Authors.
e266 Diabetic Medicine 2012 Diabetes UK
analyses were conducted to verify the primary results. HbA
1c
results were converted from per cent to mmol mol and dual
reported as follows: mmol mol (95% CI or sd); % (95% CI or
sd).
HbA
1c
targets were analysed using generalized estimating
equations for logistic regression. The 7-point self-monitored
blood glucose measures were analysed with a repeated-mea-
sures model. Models for these secondary objectives included
baseline of the measure, treatment, country, visit and the
treatment-by-visit interaction, repeated on patients across visits.
Results are expressed as mean sd unless otherwise stated.
HbA
1c
results are expressed in both mmol mol and per cent.
Results
Results from the intention-to-treat set were consistent with
those of the per protocol set.
Patient disposition and baseline characteristics
Figure 2 describes patient disposition. A total of 426 patients
were randomly assigned to treatment (LM25, n = 214; glar-
gine + lispro n = 212). Forty-two patients in the LM25 group,
compared with 26 patients in the glargine + lispro group, dis-
continued from the study. A total of 172 (80.4%) patients in
the LM25 group and 186 (87.7%) patients in the glar-
gine + lispro group completed the study (per protocol set:
LM25 n = 177; glargine + lispro n = 184). At baseline, all
patients in the full analysis set were taking metformin. Patient
demographic and baseline characteristics are described in
Table 3.
Efcacy
Glycaemic control
For the primary objective of change in HbA
1c
, from baseline to
endpoint, LM25 treatment was found to be non-inferior (upper
limit of 95% CI < 0.4) to glargine + lispro treatment, with a
least-squares mean difference in HbA
1c
(95% CI) between the
two treatment groups (LM25 glargine + lispro) of
)0.4 mmol mol (95% CI )2.7 to 1.9); )0.04% (95% CI
)0.25 to 0.17) (Fig. 3).
The least-squares mean change (95% CI) in HbA
1c
, from
baseline to endpoint for patients in the LM25 group was
)20 mmol mol (95% CI )25.9 to )14.3); )1.84% (95% CI
)2.37 to )1.31); for patients receiving glargine + lispro treat-
ment, it was )20 mmol mol (95% CI )25.5 to )13.9);
)1.80% (95% CI )2.33 to )1.27). Patients in both treatment
Enrolled (n = 616)
Randomly assigned to treatment (n = 426)
*
LM25 treatment (n = 214)
Reasons for discontinuation (n = 42)
Lost to follow-up (n = 9)
Protocol violation (n = 8)
Patient decision (n = 8)
Reasons for discontinuation (n = 26)
Lost to follow-up (n = 3)
Protocol violation (n = 8)
Patient decision (n = 4)
Physician decision (n =6)
Adverse event (n = 4)
Death (n = 2)
Entry criteria not met (n = 2)
Lack of efficacy (n = 2)
Sponsor decision (n = 1)
Physician decision (n = 3)
Adverse event (n = 5)
Death (n = 0)
Entry criteria not met (n = 2)
Lack of efficacy (n = 0)
Sponsor decision (n = 1)
Full analysis set (n = 211)

Per protocol analysis set (n = 177)

Excluded from analysis (n = 34)
Major deviation in diabetes
medications (n = 7)
Full analysis set (n = 212)
Per protocol analysis set (n = 184)
Excluded from analysis (n = 28)
Major deviation in diabetes
Completed (n = 172) Completed (n = 186)
Inclusion/exclusion criterion
violation (n = 1)
Systemic glucocorticoid therapy
(n = 0)
Early discontinuation (n = 25)
Other (n = 1)
medications (n = 15)
Inclusion/exclusion criterion
violation (n = 1)
Systemic glucocorticoid therapy
(n = 1)
Early discontinuation (n = 11)
Other (n = 0)
Glargine + lispro treatment (n = 212)
FIGURE 2 Patient disposition, number of patients. *Because of a randomization error, two patients received the opposite treatment to which they were
randomized. For the full analysis set efcacy analyses, these patients were analysed by treatment randomized to; for the per protocol analysis set and safety
analyses, these patients were analysed by the treatment received. Three patients randomized to the LM25 treatment group did not receive study
medication. These three patients were not included in the full analysis set. LM25, insulin lispro mix (25% insulin lispro, 75% insulin lispro protamine
suspension).
DIABETICMedicine Original article
2012 Eli Lilly and Company.
Diabetic Medicine 2012 Diabetes UK e267
groups experienced statistically signicant (P < 0.001) reduc-
tions in HbA
1c
from baseline to endpoint. Change in HbA
1c
over the study treatment period was similar between
treatment groups, with the least-squares mean (se) change
from baseline at week 16: LM25 )18 mmol mol (se 2.51),
)1.65% (se 0.23); glargine + lispro )19 mmol mol (se 2.51),
)1.76% (se 0.23); week 32: LM25 )21 mmol mol
(se 2.51), )1.92% (se 0.23); glargine + lispro )21 mmol mol
(se 2.51); )1.90% (se 0.23); and week 48: LM25
)21 mmol mol (se 2.51), )1.91% (se 0.23); glargine + lispro
)20 mmol mol (se 2.51); )1.87% (se 0.23).
Mean (sd) HbA
1c
at week 48 was 56 mmol mol (sd 11);
7.1% (sd 1.04) in the LM25 group and 57 mmol mol (sd 11);
7.3% (sd 1.03) in the glargine + lispro treatment group
(P = 0.741).
Mean (sd) HbA
1c
, was similar at week 48 for patients taking
one, two or three injections of LM25: 57 mmol mol (sd 11),
7.3% (sd 1.04); 55 mmol mol (sd 9), 7.2% (sd 0.82); and
57 mmol mol (sd 14), 7.3% (sd 1.26), respectively. For
patients in the glargine + lispro group, mean (sd) HbA
1c
levels
were 55 mmol mol (sd 8); 7.2% (sd 0.74) for patients taking
one glargine dose; 59 mmol mol (sd 14); 7.6% (sd 1.31) for
patients using glargine + one insulin lispro injection;
59 mmol mol (sd 14); 7.6% (sd 1.28) for patients using
glargine + two insulin lispro injections; and 56 mmol mol
(sd 10); 7.2% (sd 0.89) for patients using glargine + three
insulin lispro injections.
Over the course of the study, no statistically signicant dif-
ferences between treatment groups were found in the percent-
age of patients achieving HbA
1c
targets or postprandial blood
glucose levels. The percentage (n) of patients reaching the
< 53 mmol mol (< 7.0%) HbA
1c
target at week 48 was
40.0% (n = 68) in the LM25 group and 39.1% (n = 70) for
patients in the glargine + lispro group. For the 48 mmol mol
( 6.5%) HbA
1c
target, the percentage (n) of patients reaching
this target at week 48 was 21.2% (n = 36) and 19.0% (n = 34)
for the LM25 and glargine +lispro groups, respectively.
The least-squares mean change (se) in postprandial blood
glucose levels from baseline to study end was 5.24 mmol l
(se 1.20) for patients in the LM25 group and 4.89 mmol l
(se 1.20) for patients in the glargine + lispro group. At study
end, patients in both treatment groups had reduced 7-point
self-monitored blood glucose proles, compared with baseline
proles (Fig. 4). There were no between-treatment differences
in 7-point self-monitored blood glucose proles at baseline
Table 3 Baseline demographics and characteristics of all randomly
assigned patients
Treatment group
LM25
(n = 211)
Glargine + lispro
(n = 212)
Mean age (sd), years 56.68 (8.03) 56.29 (9.35)
Ethnicity, n (%)
African 8 (3.8) 16 (7.5)
Caucasian 69 (32.7) 61 (28.8)
East Asian 78 (37.0) 77 (36.3)
Hispanic 35 (16.6) 36 (17.0)
Native American 1 (0.5) 0 (0.0)
West Asian 20 (9.5) 22 (10.4)
Female, n (%) 118 (55.9) 107 (50.5)
Mean duration of
diabetes (sd), years
10.61 (6.18) 9.98 (5.54)
Mean body weight (sd), kg 73.84 (16.06)72.80 (13.99)
Mean BMI (sd), kg m
2
27.85 (4.11) 27.54 (4.03)
Mean HbA
1c
(sd), % 8.98 (1.04) 9.03 (1.05)
Mean HbA
1c
(sd), mmol mol 75 (11) 75 (11)
1,5-anhydroglucitol (sd), lg ml 6.24 (4.48) 6.40 (4.29)
Lipids (sd), mmol L
Total cholesterol 4.65 (1.08) 4.71 (1.20)
HDL cholesterol 1.10 (0.26) 1.10 (0.25)
LDL cholesterol 2.62 (0.83) 2.67 (1.00)
Triglycerides 2.14 (1.50) 2.12 (1.56)
Sulphonylurea use, n (%)* 208 (98.6) 210 (99.1)
LM25, insulin lispro mix (25% insulin lispro, 75% insulin
lispro protamine suspension).
*At baseline, all patients in the full analysis set were on met-
formin.
FIGURE 3 Least-squares mean change (sem) in HbA
1c
from baseline to endpoint. LM25, insulin lispro mix (25% insulin lispro, 75% insulin lispro
protamine suspension).
DIABETICMedicine Lispro 25 vs. glargine plus lispro therapy in Type 2 diabetes K. Bowering et al.
2012 The Authors.
e268 Diabetic Medicine 2012 Diabetes UK
and study end (Fig. 4). Mean change (sd) in 1,5-anhydrog-
lucitol, for the full analysis set population, from baseline to
week 48 was 5.45 lg ml (sd 6.30 lg ml) and 5.27 lg ml
(sd 5.42 lg ml) for the LM25 and gargine + lispro groups,
respectively.
Insulin dose and number of injections
The increase in insulin dose and number of injections during
the course of the study were similar for patients in both treat-
ment groups (Table 4).
Body weight and lipid proles
By study end, patients receiving LM25 treatment increased in
weight by mean (sd) 2.78 kg (sd 3.89 kg) and patients
receiving glargine + lispro treatment increased in weight by
2.92 kg (sd 3.58 kg) (Table 5). Mean change in lipid proles,
for the full analysis population, from baseline to week 48 was
similar between the two treatment groups (Table 5).
Patient evaluation of quality of life
Change from baseline to endpoint in the EQ-5D was not
statistically signicantly different between treatment groups
[between-treatment difference, least-squares mean (se), utility
score: 0.00 (se 0.02); visual analogue scale: 0.29 (se 1.44)]. For
Diabetes Health Prole (DHP-18), change from baseline to
endpoint was not statistically signicantly different between
treatment groups [between-treatment difference, least-squares
mean (se), total score: )0.85 (se 1.67); psychological distress
score: )0.20 (se 2.28); barriers to activity score: 0.90 (se 2.36);
disinhibited eating score: )3.27 (se 2.26)].
Safety
Patients in both treatment groups experienced a similar rate in
cases per 30 days [mean (sd)] of overall hypoglycaemia [LM25:
1.71 (sd 2.27); glargine + lispro: 1.96 (sd 2.77)] and nocturnal
hypoglycaemia [LM25: 0.67 (sd 1.27); glargine + lispro: 0.85
(sd 1.37)] during the study. The incidence [n (%)] of severe
hypoglycaemia was also similar between treatment groups
[LM25: 7 (3.4%); glargine + lispro: 6 (2.8%)].
Among patients in the LM25 treatment group, a similar
percentage of patients using one or two injections experienced
hypoglycaemia (56.1 and 60.8%, respectively). Among patients
using three injections, 48.5% experienced hypoglycaemia.
Similar percentages were seen in the glargine + lispro treatment
group: 56.7% in the glargine alone group; 52.4% in patients
using glargine + one insulin lispro injection; 49.0% in patients
using glargine + two insulin lispro injections, and 44.2% in
patients using glargine + three insulin lispro injections.
The number (%) of treatment-emergent adverse events [pa-
tients with 1 events: LM25, 111 (52.6%); glargine + lispro,
120 (56.6%)], and serious treatment-emergent adverse events
[patients with 1 event: LM25, 19 (9.0%); glargine + lispro,
13 (6.1%)] experienced by patients was similar between
treatment groups. Compared with patients in the glar-
gine + lispro group, patients treated with LM25 experienced
more adverse events possibly related to the study drug [patients
with 1 event: LM25, 7 (3.3%); glargine + lispro, 0 (0.0%)]
and possibly related to study device [patients with 1 event:
FIGURE 4 Seven-point self-monitored blood glucose proles at baseline
and endpoint for patients treated with LM25 or glargine + insulin lispro.
LM25, insulin lispro mix (25% insulin lispro, 75% insulin lispro prota-
mine suspension).
Table 4 Insulin dose and number of injections at week 48
Treatment group
LM25 (n = 177) Glargine + lispro (n = 184)
Mean (sd) insulin dose, U kg 0.71 (0.45) 0.71 (0.47)
Mean (sd) number of daily injections 2.14 (0.75) 2.25 (1.20)
Injection regimen, patients, n (%)
One 40 (22.6) 79 (42.9)
Two 74 (41.8) 20 (10.9)
Three 63 (35.6) 49 (26.6)
Four 36 (19.6)
LM25, insulin lispro mix (25% insulin lispro, 75% insulin lispro protamine suspension).
DIABETICMedicine Original article
2012 Eli Lilly and Company.
Diabetic Medicine 2012 Diabetes UK e269
LM25, 1 (0.5%); glargine + lispro, 0 (0.0%)]. Two patients
died as a result of myocardial infarction during the study
[LM25: 2 (0.9%); glargine + lispro: 0 (0.0%)]. Neither death
was considered related to study treatment nor delivery device
and neither patient had experienced severe hypoglycaemia
during the study.
Discussion
This is the rst randomized controlled study to investigate, in
insulin-nave patients with Type 2 diabetes, treatment with
75% insulin lispro protamine suspension, 25% insulin lispro
premixed insulin (LM25), initiated as a once-daily therapy
and with the option to be intensied up to three times a day,
compared with one daily glargine injection and progressing
up to three additional daily insulin lispro injections (glar-
gine + lispro) [911]. The results show that initiating and
intensifying insulin therapy with a low prandial ratio pre-
mixed insulin (LM25), was non-inferior to initiating and
intensifying insulin treatment with glargine + lispro, as
measured by change in HbA
1c
from baseline to study end.
Quality-of-life measures (EQ-5D and DHP-18) also indi-
cated that there were no between-treatment differences in
patient evaluation of quality of life or disease-specic quality
of life.
Compared with the proportion of patients achieving HbA
1c
targets of < 7.0% and 6.5% in a previous study of biphasic
insulin aspart 70 30 (BIAsp 30) by Garber and colleagues
[13], a smaller proportion of patients in the present study
achieved HbA
1c
targets. In this previous study, insulin regi-
mens were intensied only in patients who did not reach
glycaemic targets. Patients who met glycaemic targets in each
phase were considered completers and were not followed. In
addition, the previous study did not employ a historical
control or a treatment comparator in the study design. In
contrast, in the present study, some patients maintained
treatment with the same regimen from beginning to end of the
study.
The HbA
1c
levels observed at endpoint for both treatment
groups were consistent with previous premixed insulin studies,
such as the Treating to Target in Type 2 Diabetes Study (4T),
which investigated glycaemic control in patients over 3 years,
comparing twice-daily biphasic insulin aspart, prandial insulin
aspart three times daily and once daily basal insulin detemir
treatment regimens [16].
The results of the present study contrast with those of a
similar study by Jain and colleagues [7], comparing glycaemic
control achieved when progressing from one up to three daily
injections of 50% insulin lispro and 50% insulin lispro prot-
amine suspension (LM50), with once-daily insulin glargine plus
up to three mealtime insulin lispro injections [7]. The results of
this previous study did not support non-inferiority of LM50
therapy over glargine + lispro therapy, based on the pre-spec-
ied non-inferiority margin of 0.3%; no statistically signicant
difference between treatment groups in HbA
1c
reduction from
baseline to endpoint were reported. The previous study, how-
ever, did not allow addition of insulin injections after 24 weeks
[7]. Interestingly, even although physicians in the present study
were allowed to increase the number of injections beyond
24 weeks, the majority of patients in both studies were using
one or two injections per day. No data were captured in the
present study to elucidate reasons for lack of progressive
increase in the number of injections, so no further conclusions
can be made.
InthestudybyJainandcolleagues, patients inthe LM50group
could be switched to an evening dose of LM25. Lower HbA
1c
values were seen in some patients after switching, which may
suggest that LM25, with a higher proportion of basal insulin,
might have contributed to the improved glycaemic control that
was found in those patients [7]. However, as the present study
design did not include a third LM50 treatment group, we cannot
conclude this hypothesis with the current results.
A study by Rosenstock and colleagues evaluated LM50 given
as three daily injections compared with bedtime glargine plus
insulin lispro at mealtime [12]. In this previous study, 54% of
patients taking LM50 achieved the target HbA
1c
of 7.0%,
compared with 69% of patients in the glargine + lispro group
(P = 0.009). This contrasts with the present study, in which
40% of patients treated with LM25 achieved the target HbA
1c
.
Baseline HbA
1c
was similar in both studies. However, there
were important differences in these two studies. In the present
study, patients were nave to insulin therapy and were failing
on oral treatment. In the study by Rosenstock and colleagues,
patients were already on insulin. The duration of diabetes was
longer and the mean total daily glargine dose at study entry was
more than 50 units per day. In addition, three mealtime
injections of LM50 were used from the start, rather than fol-
lowing a regimen of progressive intensication [12]. These
results may suggest that starting a more intense regimen right
away may produce greater reduction in HbA
1c
than progressive
intensication. However, without a head-to-head comparison,
it is not feasible to draw conclusions regarding the efcacy of
one treatment over another.
Table 5 Body weight and lipids at study end
Treatment group
LM25
(n = 211)
Glargine + lispro
(n = 212)
Mean (sd) weight increase
from baseline, kg
2.78 (3.89) 2.92 (3.58)
Mean change (sd) in lipid proles from baseline, mmol l
Cholesterol )0.11 (0.84) )0.09 (0.82)
LDL cholesterol )0.04 (0.73) )0.07 (0.70)
HDL cholesterol 0.07 (0.17) 0.03 (0.18)
Triglycerides )0.37 (1.14) )0.19 (1.41)
LM25, insulin lispro mix (25% insulin lispro, 75% insulin
lispro protamine suspension).
DIABETICMedicine Lispro 25 vs. glargine plus lispro therapy in Type 2 diabetes K. Bowering et al.
2012 The Authors.
e270 Diabetic Medicine 2012 Diabetes UK
An ongoing study (Clinical Trials Registry No; NCT
01175811) comparing LM50 twice daily (breakfast, lunch) and
LM25 (dinner) with once-daily insulin glargine and three pre-
meal insulin lispro injections may enhance our understanding
of the importance of intensifying mealtime insulin in Type 2
diabetes.
The mean number of injections and total daily mean insulin
dose administered by patients over the course of the study and
at endpoint was similar between LM25 and glargine + lispro
treatment groups, with patients administering approximately
two daily injections. While the number of injections in the
study by Jain and colleagues [7] is consistent with the present
study, with approximately two injections at endpoint for both
the LM50 and basalbolus treatment groups, the mean number
of insulin injections and weight-adjusted total daily insulin
dosage were reported to be signicantly greater in the LM50
group compared with the basalbolus group [7]. This differ-
ence could reect the increased need for basal coverage in
patients treated with LM50, compared with the present study
of LM25.
In the present study, approximately 43% of patients were
taking only one injection of glargine, perhaps attributable to
the lack of forced titration. Rather than following a predeter-
mined regimen of adding injections, our study allowed the
investigator to make the determination to intensify treatment
based on the patients blood glucose values. These results may
signal reluctance of some investigators to intensify treatment
unless intensication is forced by the protocol. We did not
include measurements of investigator compliance with the
titration algorithm, however, but we controlled by monitoring
and by capturing protocol violations; major protocol violations
were excluded from analyses.
Patients in both treatment groups experienced a similar rate
and incidence of overall and nocturnal hypoglycaemia; the
incidence of severe hypoglycaemia was also similar between
treatment groups. Although there were numerically more
treatment-emergent adverse events related to study drug, device
or procedure experienced by patients in the LM25 group
compared with the glargine + lispro group, the difference was
not statistically signicant. As there were a small number of
patients who experienced these events, it is not possible to
assign causality, and it is reasonable that these are likely
attributable to random variation in the study population.
Overall, the safety prole for patients treated with the LM25
regimen was similar to that for patients treated with the glar-
gine + lispro regimen. The safety prole for patients in this
study is consistent with previous premixed insulin studies
[7,13,16].
A difference in treatment discontinuation rates was observed,
with more discontinuations based on patient and physician
decisions and patients lost to follow-up reported in the LM25
treatment group compared with the glargine + lispro treatment
group. There is no indication in the data as to the cause of this
disparity; thus, an attempt to attribute the result to a specic
cause would be mainly speculative.
Limitations of this study include the open-label design,
because of the different appearance of the insulin preparations,
and that insulin titration was not a forced process and therefore
open to local variation in treatment intensication practices.
Conclusion
This study has shown that, in insulin-nave patients with
Type 2 diabetes inadequately controlled with oral anti-hy-
perglycaemic medications, glycaemic control achieved when
initiating and intensifying insulin treatment with LM25
premixed insulin is non-inferior to that achieved when initi-
ating and intensifying insulin treatment with glargine + lis-
pro. These ndings may assist physicians in providing
patients with additional options when initiating and intensi-
fying insulin treatment following oral anti-hyperglycaemic
medication failure; however, further studies are warranted to
provide more clarity regarding premixed analogue insulin
regimens.
Competing interests
This study was funded by Eli Lilly. JO, JL and VAR are
employees of Eli Lilly. JO and JL are shareowners in Eli Lilly.
JF, KB and LJ were investigators on the F3Z-CR-IOPH study.
JF has no competing interests to declare. KB has been reim-
bursed by Eli Lilly for attending symposia, received honoraria
from Eli Lilly and has served on advisory boards for Eli Lilly.
LJ has received consulting and lecture fees from Eli Lilly,
Bristol-Myers Squibb, Novartis, Novo Nordisk, Merck, Bayer,
Takeda, Sano-Aventis, GlaxoSmithKline, Roche, Johnson &
Johnson, Boehringer Ingelheim and Guangzhou Zhongyi
Pharmaceutical, and has received grants from Guangzhou
Zhongyi Pharmaceutical and Bayer.
Acknowledgements
This study was funded by Eli Lilly and Company. The
authors are indebted to the patients who participated in this
study and their families, and to all of the physicians, nurses
and study coordinators who cared for the patients. The
authors would like to thank the F3Z-CR-IOPH study inves-
tigators and study team; Alexander Simpson for assistance
with study conception and design; Cherry He and Yueshan
Feng for statistical support; and John Chan and Yao Lei for
critically reviewing the manuscript for important intellectual
content.
References
1 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: 15771589.
2 Turnbull FM, Abraira C, Anderson RJ, Byington RP, Chalmers JP,
Duckworth WC et al. Intensive glucose control and macrovascular
outcomes in type 2 diabetes. Diabetologia 2009; 52: 22882298.
DIABETICMedicine Original article
2012 Eli Lilly and Company.
Diabetic Medicine 2012 Diabetes UK e271
3 Nathan DM, Buse JB, Davidson MB, Ferrannini E, Holman RR,
Sherwin R et al. Medical management of hyperglycemia in
type 2 diabetes: a consensus algorithm for the initiation and
adjustment of therapy: a consensus statement of the American
Diabetes Association and the European Association for the Study
of Diabetes. Diabetes Care 2009; 32: 193203.
4 Rodbard HW, Jellinger PS, Davidson JA, Einhorn D, Garber AJ,
Grunberger G et al. Statement by an American Association of
Clinical Endocrinologists American College of Endocrinology
consensus panel on type 2 diabetes mellitus: an algorithm for
glycemic control. Endocr Pract 2009; 15: 540559.
5 Garber AJ, Ligthelm R, Christiansen JS, Liebl A. Premixed insulin
treatment for type 2 diabetes: analogue or human? Diabetes Obes
Metab 2007; 9: 630639.
6 Liebl A, Prager R, Binz K, Kaiser M, Bergenstal R, Gallwitz B,
PREFER Study Group. Comparison of insulin analogue regimens
in people with type 2 diabetes mellitus in the PREFER Study: a
randomized controlled trial. Diabetes Obes Metab 2009; 11: 45
52.
7 Jain SM, Mao X, Escalante-Pulido M, Vorokhobina N, Lopez I,
Ilag LL. Prandial-basal insulin regimens plus oral antihyperglycae-
mic agents to improve mealtime glycaemia: initiate and progres-
sively advance insulin therapy in type 2 diabetes. Diabetes Obes
Metab 2010; 12: 967975.
8 Ligthelm RJ, Mouritzen U, Lynggaard H, Landin-Olsson M, Fox C,
le Devehat C et al. Biphasic insulin aspart given thrice daily is as
efcacious as a basalbolus insulin regimen with four daily injec-
tions: a randomised open-label parallel group four months com-
parison in patients with type 2 diabetes. Exp Clin Endocrinol
Diabetes 2006; 114: 511519.
9 Ilag LL, Kerr L, Malone JK, Tan MH. Prandial premixed insulin
analogue regimens versus basal insulin analogue regimens in the
management of type 2 diabetes: an evidence-based comparison.
Clin Ther 2007; 29: 12541270.
10 Christiansen JS, Liebl A, Davidson JA, Ligthelm RJ, Halimi S. Mid-
and high-ratio premix insulin analogues: potential treatment options
for patients withtype 2diabetes inneedof greater postprandial blood
glucose control. Diabetes Obes Metab 2010; 12: 105114.
11 Qayyum R, Bolen S, Maruthur N, Feldman L, Wilson LM, Mar-
inopoulos SS et al. Systematic review: comparative effectiveness and
safety of premixed insulin analogues in type 2 diabetes. Ann Intern
Med 2008; 149: 549559.
12 Rosenstock J, Ahmann AJ, Colon G, Scism-Bacon J, Jiang H,
Martin S. Advancing insulin therapy in type 2 diabetes previously
treated with glargine plus oral agents: prandial premixed (insulin
lispro protamine suspension lispro) versus basal bolus (glar-
gine lispro) therapy. Diabetes Care 2008; 31: 2025.
13 Garber AJ, Wahlen J, Wahl T, Bressler P, Braceras R, Allen E. et al.
Attainment of glycaemic goals in type 2 diabetes with once-, twice-,
or thrice-daily dosing with biphasic insulin aspart 70 30 (The 1-2-3
study). Diabetes Obes Metab 2006; 8: 5866.
14 World Medical Association Declaration of Helsinki. Recommen-
dations guiding physicians in biomedical research involving human
subjects. J Am Med Assoc 1997; 277: 925926.
15 WHO. Denition and Diagnosis of Diabetes Mellitus and Inter-
mediate Hyperglycemia: Report of a WHO IDF Consultation.
Geneva: World Health Organization, 2006.
16 Holman RR, Farmer AJ, Davies MJ, Levy JC, Darbyshire JL,
Keenan JF et al. Three-year efcacy of complex insulin regimens in
type 2 diabetes. N Engl J Med 2009; 361: 17361747.
DIABETICMedicine Lispro 25 vs. glargine plus lispro therapy in Type 2 diabetes K. Bowering et al.
2012 The Authors.
e272 Diabetic Medicine 2012 Diabetes UK