O R I G I NALAR T I C L E doi: 10.1111/j.

1463–
1326.2004.00373.x

Lower within-subject variability of fasting blood glucose and

reduced weight gain with insulin detemir compared to NPH
insulin in patients with type 2 diabetes

T. Haak,1A. Tiengo,2E. Draeger,3M. Suntum3and W. Waldha¨usl4

1Research Institute of the Diabetes Academy Mergentheim, Bad Mergentheim, Germany
2Department of Clinical and Experimental Medicine, University of Padova Medical School, Padova,
Italy
3Novo Nordisk A/S, Gladsaxe, Denmark
4Department of Medicine III, Medical University of Vienna, Vienna, Austria

Aim:The aim of this study was to compare the efficacy and safety of a basal-bolus insulin regimen
comprising either insulin detemir or neural protamine hagedorn (NPH) insulin in combination with
mealtime insulin aspart in patients with type 2 diabetes. Tujuan dari penelitian ini adalah untuk
membandingkan efektivitas dan keamanan regimen insulin basal bolus antara insulin detemir
atau insulin neural protamine hagedorn (NPH) yang dikombinasikan dengan insulin aspart
mealtime pada pasien DM tipe 2.
Methods: This was a 26-week, multinational, open-label, parallel group trial with 505 patients with type
2 diabetes (mean age, 60.4 8.6years; mean BMI, 30.4 5.3kg/m2; mean HbA1c, 7.9 1.3%). Patients,
randomized 2:1 to insulin detemir or NPH insulin, received basal insulin either once or twice daily
according to their pretrial insulin treatment and insulin aspart at mealtimes. Penelitian ini dilakukan
selama 26 minggu, multinasional, open label, kelompok parallel dengan 505 pasien dengan DM
tipe 2 (usia rata-rata: 60,4±8,6 tahun; BMI rata-rata: 30,4±5,3 kg/m2; HbA1c rata-rata:
7,9±1,3%). Pasien diacak dengan perbandingan 2:1 untuk insulin detemir atau insulin NPH,
menerima insulin basal 1 atau 2 kali sehari tergantung pada treatment insulin yang mereka
dapatkan sebelum penelitian dilakukan dan insulin aspart pada waktu makan.
Results: After 26weeks of treatment, significant reductions in HbA1c were observed for insulin detemir
(0.2%-points, p¼0.004) and NPH insulin (0.4%-points; p¼0.0001); HbA1c levels were comparable at
study end (insulin detemir, 7.6%; NPH insulin, 7.5%). The number of basal insulin injections
administered per day had no effect on HbA1c levels (p¼0.50). Nine-point self-measured blood glucose
(SMBG) profiles were similar for the two treatment groups (p¼0.58), as were reductions in fasting
plasma glucose (FPG) (insulin detemir, 0.5mmol/l; NPH insulin, 0.6mmol/l). At study end, FPG
concentrations were similar for the two treatment groups (p¼0.66). By contrast, within-subject day-to-
day variation in fasting SMBG was significantly lower with insulin detemir (p¼0.021). Moreover,
patients receiving insulin detemir gained significantly less body weight than those who were
administered NPH insulin (1.0 and 1.8kg, respectively, p¼0.017). The frequency ofadverse eventsand
the risk ofhypoglycaemia werecomparablefor the two treatment groups. Setelah 26 minggu
pengobatan, terjadi penurunan nilai HbA1c secara signifikan pada pemberian insulin detemir
(0,2% point, p=0,004) dan insulin NPH (0,4% point; p=0,0001); level HbA1c dibandingkan pada
akhir penelitian (insulin detemir 7,6%; insulin NPH 7,5%). Frekuensi injeksi insulin yang
diadministrasikan per harinya tidak memiliki efek terhadap kadar HbA1c (p=0,50). Profil
Sembilan poin kadar gula darah yang diukur sendiri (SBGM) mirip pada 2 kelompok subjek
penelitian (p=0,58) sama halnya dengan penurunan kadar glukosa darah puasa (FPG) (insulin
detemir 0,5 mmol/L; insulin NPH 0,6 mmol/L). Pada akhir penelitian, 2 kelompok subjek
penelitian menunujukkan nilai konsentrasi FPG yang hampir sama (p=0,66). Sebaliknya, subjek
yang mendapatkan insulin detemir menunjukkan penurunan fasting SBMG yang signifikan
(p=0,021). Sementara itu, pasien yang menerima insulin detemir juga menunjukkan penurunan
berat badan daripada kelompok subjek yang menerima insulin NPH (1.0 dan 1.8 kg, respectively,
p=0,017). Frekuensi efek samping dan risiko hipoglikemia juga dibandingkan antara 2 kelompok
subjek penelitian.
Conclusions: Patients with type 2 diabetes, treated for 26weeks with insulin detemir plus insulin aspart
at mealtimes, experienced comparable glycaemic control but significantly lower within-subject
variability and less weight gain compared to patients treated with NPH insulin and insulin aspart.
Insulin detemir was well tolerated and had a similar safety profile to NPH insulin. Pasien dengan DM
tipe 2 yang ditreatment selama 26 minggu dengan insulin detemir dan insulin aspart pada waktu
makan dibandingkan dengan kelompok kontrol. Kelompok subjek penelitian yang menerima
insulin detemir dan insulin aspart secara signifikan menunjukkan kadar glukosa darah yang
lebih rendah. Selain itu, kelompok subjek penelitian dengan insulin detemir mengalami
penurunan berat badan dibandingkan dengan insulin NPH dan insulin aspart. Insdulin detemir
dapat ditoleransi dengan baik dan memiliki profil keamanan yang hampir sama dengan insulin
NPH.
Keywords: basal-bolus, body weight, insulin detemir, type 2 diabetes, variability
Received 19 December 2003; returned for revision 10 March 2004; revised version accepted 12 March
2004

Introduction
Type 2 diabetes is a chronic metabolic disorder, currently affecting 140–150 million people worldwide,
which often leads to long-term microvascular and macrovascular complications that substantially
increase morbidity and mortality. DM tipe 2 adalah kelainan metabolic kronis yang diderita sekitar
140-150 juta orang di seluruh dunia. Pada jangka panjang dapat mengakibatkan komplikasi
mikrovaskular dan makrovaskular yang dapat meningkatkan morbiditas dan mortalitas.
Traditionally, oral antidiabetic drugs (OADs) constitute the first line of treatment for type 2 diabetes
when diet and exercise fail to provide satisfactory glycaemic control. However, as the disease advances,
progressive loss of pancreatic b-cell function necessitates the use of insulin therapy when glycaemic
targets cannot be achieved by oral therapy alone. Secara konvensional, Obat oral antidiabetes (OAD)
merupakan terapi lini pertama pada DM tipe 2 ketika diet dan olahraga gagal mencapai kadar
gula darah normal. Namun demikian, peningkatan penyakit ini dapat menyebabkan peningkatan
progresitivas kerusakan fungsi sel beta pancreas sehingga dibutuhkan terapi insulin ketika kadar
gula darah normal tidak dapat tercapai dengan terapi oral saja, Although insulin has commonly
been regarded as a final treatment option, several recent studies have provided evidence that tight
glycaemic control, achieved in part through earlier and more intensive use of insulin, can reduce the
incidence and delay the progression of a number of long-term complications associated with type 2
diabetes. Moreover, evidence suggests that in contrast to sulphon- ylureas, early insulin treatment in
newly diagnosed patients with type 2 diabetes may temporarily prolong endogenous insulin secretion
and promote better metabolic control. Meskipun terapi insulin umumnya digunakan sebagai pilihan
terapi final, namun dari beberapa penelitian menunjukkan bahwa penggunaan insulin lebih awal
dengan lebih intensif dapat menurunkan insiden dan memperlambat progresifitas komplikasi
jangka panjang yang berhubungan dengan DM tipe 2. Usually, insulin therapy is initiated as a daily
injection of long-acting basal insulin in combination with OADs. However, when glycaemic control
becomes inadequate, a basal-bolus insulin regimen is often adopted. Although the development of
short-acting bolus insulin analogues (e.g. insulin aspart and insulin lispro) has improved postprandial
glycaemic control, traditional basal insulin preparations [e.g. neural prot- amine hagedorn (NPH)
insulin and Ultralente1 insulin] cannot deliver insulin at the constant and reproducible low level that
characterizes normal insulin secretion. Biasanya, terapi insulin dimulai dengan injeksi insulin basal
long acting sehari sekali yang dikombinasikan dengan terapi oral. Namun, jika kadar gula darah
tidak adekuat (belum mencapai kadar normal), maka digunakan regimen insulin basal bolus.
Meskipun analog short acting bolus insulin (e.g insulin aspart dan insulin lispro) dapat
meningkatkan kontrol gula darah postprandial (gula darah setelah makan), insulin basal
konvensional (insulin NPH dan ultralente) tidak dapat mendistribusikan insulin secara konstan
dan diproduksi dalam kadar rendah yang mengkarakterisasi sekresi insulin normal. For example,
NPH insulin reaches a peak in plasma concentration approximately 5h after injection and has an
insufficient duration of action to cover night-time requirements. Furthermore, inadequate resuspension
prior to injection and variations in crystal size make dosing precision and absorption kinetics highly
variable, resulting in unpredictable glucose levels. Sebagai contoh, insulin NPH mencapai kadar
puncak konsentrasi plasma sekitar 5 jam setelah diinjeksikan dengan durasi kerja yang tidak
mencukupi untuk mengcover kebutuhan insulin pada malam hari. Selain itu, resuspensi injeksi
yang tidak adekuat serta variasi ukuran kristal menyebabkan ketepatan dosis dan kinetika
absorbs nya sangat bervariasi menghasilkan kadar glukosa darah yang tidak dapat diprediksi.
These limitations have therefore prompted the development of new basal insulin analogues (insulin
glargine and insulin detemir of which insulin detemir, a derivative of human insulin soluble at neutral
pH, is the most recent to be developed. The protracted action of insulin detemir largely stems from
delayed absorption from the injection site due to self- association and binding, via a covalently attached
fatty acid moiety, to albumin in the interstitial fluid. This novel mechanism of protraction provides more
reproducible insulin absorption and lower within-subject variability than other basal insulins.
Keterbatasan ini menyebabkan peningkatan perkembangan analog insulin basal yang baru
(insulin glargine dan insulin detemir, dimana insulin detemir merupakan derivate insulin
manusia yang larut pada pH netral dan pada akhir-akhir ini banyak dikembangkan). Insulin
detemir diabsorbsi secara perlahan (diperlama) dari tempat injeksi bergantung pada ikatan
kovalen terhadap asam lemak dan albumin di cairan intersisial. Mekanisme ini dapat
meningkatkan absorbsi insulin dan menurunkan factor variabilitas individu dibandingkan
insulin basal lainnya. Comparative studies of insulin detemir and NPH insu- lin in patients with type 1
diabetes have demonstrated similar glycaemic control but a more extended and smoother time-action
profile with insulin detemir, properties that are associated with lower within-subject variability in
fasting blood glucose and a reduced risk of hypoglycaemia. Penelitian yang membandingkan antara
antara insulin detemir dan insulin NPH pada DM tipe 1 menunjukkan profil glukosa darah yang
hampir sama namun durasi kerja dari insulin detemir lebih panjang. Durasi kerja yang lebih
panjang ini membuat insulin detemir memliki risiko efek hipoglikemi yang lebih rendah serta
dapat menurunkan factor variabilitas individu. Furthermore, compared to NPH insulin, patients
receiving insulin detemir gain signifi- cantly less body weight at comparable metabolic control. These
characteristics could be particu- larly advantageous in the treatment of type 2 diabetes where
approximately 80% of newly diagnosed patients are overweight or obese. Selanjutnya, dibandingkan
dengan pasien yang yang menerima insulin NPH, pasien dengan insulin detemir mengalami
penurunan berat badan yang lebih signifikan. Hal ini dapat menjadi keuntungan pada pasien
tipe 2 yang 80% diantaranya mengalami kelebihan berat badan atau obesitas.For patients with
type 2 diabetes, fears of weight gain and hypoglycaemia asso- ciated with intensive insulin therapy have
often resulted in a reluctance to initiate insulin treatment and/or titrate insulin to a dose where near-
normal gly- caemia is achieved. The present study was conducted to compare the effi- cacy, safety and
weight change associated with a basal- bolus regimen comprising either insulin detemir or NPH insulin
in combination with mealtime insulin aspart in patients with type 2 diabetes previously treated with
insulin. Untuk pasien DM tipe 2, ketakutan peningkatan berat badan dan kejadian hipoglikemia
yang berhubungan dengan terapi insulin yang terlalu intensif membuat pasien enggan memulai
terapi insulin atau titrasi dosis insulin mendekati dosis yang dapat mencapai kadar glukosa
darah normal. Penelitian ini membandingkan efektifitas, keamanan dan perubahan berat badan
yang berhubungan dengan regimen insulin basal bolus menggunakan insulin detemir atau insulin
NPH yang dikombinasikan dengan insulin aspart pada waktu makan pada pasien DM tipe 2 yang
sebelumnya diterapi dengan insulin.

Research Design and Methods

Sixty-three sites in five European countries participated in this open-label, parallel-group trial. The trial
con- sisted of a 3-week screening period and a 26-week treat- ment period. The study was approved by
local ethics committees and health authorities according to local regulations and was conducted in
accordance with ICH-Good Clinical Practice guidelines and the Declaration of Helsinki. Signed
informed consent was provided before any trial-related activities. 63 lokasi di 5 negara di Eropa
berpartisipasi dalam penelitian open label, parallel group ini. Penelitian ini terdiri dari periode
screening selama 3 minggu dan treatment selama 26 minggu.Penelitian ini disetujui oleh komite
etik lokal dan otoritas kesehatan sesuai dengan peraturan lokal dan dilakukan sesuai dengan
pedoman ICH-Good Clinical Practice dan Deklarasi Helsinki. Inform consent yang sudah
ditandatangani telah tersedia sebelum penelitian dilakukan.
Patients
A total of 505 patients with type 2 diabetes of 12months duration, aged 35years, HbA1c 12.0% and who
had received insulin treatment for at least 2months (basal insulin dose 30% of the total daily insulin
dose) were included in the trial. Patients were excluded if they had received OADs within 2months of
the trial, were pregnant or breast feeding, suffered from proliferative retinopathy, uncontrolled
hypertension, recurrent major hypoglycaemia, impaired renal or hepatic function, cardiac problems, or
if they received a total daily basal insulin dose >100IU/day. Total pasien pada penelitian ini adalah
505 pasien dengan DM tipe 2 selama ≥12 tahun, usia ≥35 tahun, HbA1c ≤12% dan menerima
terapi insulin minimal selama 2 bulan (dosis insulin basal ≥30% dati total dosis insulin sehari).
Pasien diekslusi jika menerima terapi antisiabetes oral dalam 2 bulan masa penelitian, pasien
hamil atau menyusui, menderita proliferative retinopati, hipertensi tidak terkontrol, sering
mengalami hipoglikemia, mengalami gangguan fungsi ginjal dan hepar, memiliki masalah
jantung atau jika mereka menerima insulin basal dengan total dosis sehari >1000 IU/hari.
Study Design
Patients were randomized (2:1) to receive either insulin detemir (LevemirTM, Novo Nordisk A/S;
100U/ml) or NPH insulin (isophane human insulin, Novo Nordisk A/S, Copenhagen, Denmark;
100IU/ml) as basal insulin for 26weeks. Patients receiving more than one basal insulin injection per day
prior to the trial received twice-daily (before breakfast and at bedtime) injections following
randomization. All other patients, including those receiving biphasic (premixed insulin) prior to the
trial, received a once-daily basal insulin injection (at bedtime). For all patients, insulin aspart
(NovoRapid1, Novo Nordisk A/S, 100U/ml) was administered imme- diately prior to each main meal.
All insulin preparations were injected subcutaneously (basal insulin, abdomen or thigh; bolus insulin,
abdomen) using a NovoPen1 3 device (Novo Nordisk A/S). Throughout the trial, basal insulin doses
were optimized according to self-mea- sured blood glucose (SMBG) values. Patients rando- mized to
insulin detemir were initiated on 50% of their pretrial basal insulin dose whereas those randomized to
NPH insulin continued with their pretrial basal insulin dose. During the trial, patients were instructed to
aim for the following SMBG levels: pre breakfast, 4–7mmol/l; nocturnal (02:00–04:00), 4–7mmol/l;
postprandial (90min after meal), <10mmol/l. If a patient did not reach these SMBG targets on a once-
daily basal insulin regimen (either with insulin detemir or NPH insulin), and further dose escalations
were inadvisable, they were transferred to a twice-daily basal insulin regimen with appropriate dose
adjustments. Pasien diacak 2:1 untuk menerima insulin detemir (LevemirTM, Novo Nordisk A/S;
100U/ml) atau insulin NPH (isophane human insulin, Novo Nordisk A/S, Copenhagen, Denmark;
100IU/ml) sebagai insulin basal selama 26 minggu. Pasien yang menggunakan injeksi insulin
basal ebih dari sekali sehari akan diberikan dosis 2 kali sehari (sebelum sarapan dan pada waktu
menjelang tidur) selama penelitian. Sedangkan pasien yang lain, termasuk pasien yang biasa
menerima insulin premix akan diberikan dosis injeksi insulin satu kali sehari pada waktu
menjelang tidur. Untuk semua pasien, insulin aspart (NovoRapid1, Novo Nordisk A/S, 100U/ml)
diadministrasikan segera pada saat makan. Semua insulin diinjeksikan secara subkutan (basal
insulin, abdomen or thigh; bolus insulin, abdomen) menggunakan NovoPen 13 device (Novo
Nordisk A/S). Selama penelitian, dosis insulin basal dioptimasikan sesuai dengan hasil
pengukuran SBMG. Pasien yang menerima insulin detemir dimulai dengan pemberian dosis 50%
dari dosis insulin basal yang biasa mereka gunakan. Sementara pasien yang menerima insulin
NPH diberikan dosis sesuai dengan dosis insulin basal awal (sebelum penelitian) yang mereka
gunakan. Selama penelitian, pasien diharapkan dapat mencapai target yang telah ditentukan.
Yaitu SMBG levels: pre breakfast, 4–7mmol/l; nocturnal (02:00–04:00), 4–7mmol/l; postprandial
(90min after meal), <10mmol/l. Jika pasien tidak mencapai target tersebut, maka dosis akan
diganti menjadi 2 kali sehari disertai dengan penyesuaian dosis.
Methods
HbA1c (reference range of assay: 4.3–6.1%) was deter- mined in whole blood by high-performance
liquid chro- matography (HPLC) using a Bio-Rad VariantTM instrument. Clinic fasting plasma glucose
(FPG) concen- trations were determined by a hexokinase method. These and standard safety
determinations (haematology and biochemistry) were performed by a central labora- tory
(Laboratorium fu¨r Klinische Forschung GmbH, Raisdorf, Germany). One Touch Profile1 blood
glucose meters (LifeScan, Neckargemu¨nd, Germany) calibrated for blood glucose measurements were
used for all SMBG assessments. HbA1c and FPG were measured at the screening visit and after 13 and
26weeks of treat- ment. Patients performed 9-point SMBG profiles before randomization and after 13
and 26weeks of treatment. In addition, fasting SMBG levels were measured on each of the final 7days
of treatment. Adverse events, coded according to the Novo Nordisk Adverse Reaction Dictionary, were
monitored through- out the study period and, where necessary, were treated by established methods of
care. Treatment emergent adverse events were defined as events occurring from
the first day of dosing to 7days following the final dose. Clinical laboratory assessments (haematology
and bio- chemistry), physical examinations, fundoscopy/fundus photography,12-
leadelectrocardiogram(ECG),andmeas- urements of vital signs (blood pressure and pulse) and body
weight were performed at screening, randomiza- tion and after 13 and 26weeks of treatment (as appro-
priate). Hypoglycaemic episodes were classified as major when a subject was unable to treat
him/herself, as minor if blood glucose concentration was <2.8mmol/l but no assistance was required,
and as symptoms only if the episode was not confirmed by a blood glucose meas- urement. Nocturnal
hypoglycaemia was defined as any hypoglycaemic episode that occurred between the hours of 23:00
and 06:00. HbA1c (N: 4.3-6.1%) diukur pada whole blood dengan HPLC menggunakan
instrument Bio-Rad Varian TM. Konsentrasi FPG diukur menggunakan metode hexokinase.
Standart keamanan dilakukan oleh laboratorium sentral (Laboratorium fu¨r Klinische
Forschung GmbH, Raisdorf, Germany). Pengukuran SBMG menggunakan One Touch
Glucometer yang sudah dikalibrasi. HbA1c dan FPG diukur pada kunjungan saat dilakukan
screening, setelah 13 minggu dan 26 minggu terapi. Terdapat 9 point profil SBMG yang diukur
pada pasien sebelum dilakukan randomisasi, setelah 13 minggu dan 26 minggu terapi. Kadar
SBMG puasa juga diukur setiap 7 hari terapi. Efek samping, mengacu pada Novo Nordisk
Adverse Reaction Dictionary dimonitor selama penelitian dilakukan, jika terjadi efek samping
maka ditreatment sesuai dengan metoda penanganan yang tepat.penanganan emergency akibat
efek samping didefinisikan sebagai kejadian yang terjadi mulai hari pertama dosis diberikan
sampai 7 hari berikutnya dengan dosis final. Penilaian laboratorium klinis (hematologi dan bio-
kimia), pemeriksaan fisik, funduskopi / fundus fotografi,12-leadelectrocardiogram (EKG), dan
pengukuran tanda-tanda vital (tekanan darah dan denyut nadi) serta berat badan dilakukan
pada saat skrining, randomisasi, dan setelah 13 dan 26 minggu pengobatan. Episode
hipoglikemik diklasifikasikan sebagai major ketika subjek tidak dapat mengatasinya sendiri,
sebagai minor jika konsentrasi glukosa darah <2.8mmol / l, tetapi tidak perlu terapi, dan sebagai
gejala hanya jika kadar gula darah ternyata dalam rentang normal. episode itu tidak
dikonfirmasi oleh meas glukosa darah - urement. Hipoglikemia nokturnal didefinisikan sebagai
hipoglikemia yang terjadi antara jam 23:00 dan 06:00

Statistical Analyses
All analyses were based on the intention-to-treat (ITT) analysis set, which included all patients
randomized and exposed to at least one dose of study drug. The primary endpoint, HbA1c (%) after
26weeks of treat- ment, was evaluated by an analysis of variance (ANOVA) model, with treatment and
country as fixed effects and covariate adjustment for baseline values. The study had sufficient power
(85%) to detect an average difference of 0.4% in HbA1c between treatment groups. A 95% two- sided
confidence interval (CI) was constructed for the difference between the treatment group means (insulin
detemirNPH insulin); insulin detemir was deemed non-inferior if the upper limit of the 95% CI was
<0.4% (absolute). Treatments were considered compar- able if the non-inferiority criterion was
fulfilled. The influence of the number of basal insulin injec- tions per day on HbA1c levels following
26weeks of treatment was investigated by including the number of basal injections as a factor in the
ANOVA model used for the primary endpoint while successively including and excluding the treatment
by number of basal injections interaction term. Change in body weight (both with and without adjust-
ment for change in HbA1c concentration) and FPG levels following 26weeks of treatment were
analysed using a similar ANOVA model to that used for the primary end- point. The 9-point SMBG
profiles, measured after 26weeks of treatment, were analysed using a repeated measures ANOVA
model. Within-subject day-to-day varia- tion in fasting SMBG levels during the last 7days of treatment
was compared between the two treatment groups using variance component models. To estimate the
relative risk of hypoglycaemia, all hypoglycaemic episodes occurring during the maintenance period
were analysed as recurrent events using a gamma frailty model with treatment group as covariate.
Nocturnal hypoglycaemic episodes were analysed separately (as above). All results are presented as
means SD unlessotherwisespecified.Thesignificancelevelwas5%. Semua analisis didasarkan pada
perangkat analisis Intention To Treat (ITT), yang mencakup semua pasien secara acak. Hasil
pengukuran HbA1c setelah 26 minggu erapi dievaluasi menggunakan analisis varians (ANOVA).
Penelitian ini memiliki power 85% untuk mendeteksi perbedaan rata-rata 0,4% pada HbA1c
antar kelompok perlakuan. Interval kepercayaan (CI) two sided 95% digunakan untuk melihat
perbedaan rata-rata antar kelompok perlakuan (insulin detemir dan insulin NPH); Insulin
detemir dianggap noninferior jika batas atas 95% CI adalah <0,4% (absolut). Pengobatan dapat
dibandingkan jika semua criteria noninferior dapat terpenuhi. Pengaruh frekuensi injeksi insulin
basal per hari terhadap tingkat HbA1c setelah 26 minggu terapi dievaluasi dengan memasukkan
frekuensi injeksi insulin basal sebagai factor dalam model ANOVA yang digunakan sebagai titik
akhir primer. Kemudian secara berturut turut mengganti frekuensi terapi injeksi insulin sesuai
dengan kelompok subjek. Perubahan pada berat badan (baik dengan atau tanpa perubahan
konsentrasi HbA1c) dan kadar FPG setalah 26 minggu terapi dianalisa menggunakan model
ANOVA yang serupa seperti halnya model ANOVA yang digunakan pada pengukutan titik akhir
primer. Sementara profil 9 point SBMG yang diukur setelah 26 minggu terapi dianalisa
menggunakan model pengukuran berulang ANOVA. Variasi kadar SBMG puasa subjek dari hari
ke hari selama 7 hari terakhir penelitian dibandingkan dengan 2 kelompok subjek menggunakan
variance component model. Untuk mengestimasi/ memperkirakan risiko relative hipoglikemia
(Relative Risk), semua kejadian hipoglikemia yang terjadi selama periode maintenance dianalisa
sebagai kejadian berulang menggunakan model gamma frailty dengan kelompok terapi sebagai
covariate. Episode hipoglikemia nocturnal dianalisa secara terpisah. Semua hasil nya disajikan
sebagai simpangan rata-rata (means SD). Significant level nya 5%.

Results
Overall, the two treatment groups were evenly matched with respect to demographic and baseline
characteristics (table1). Of those receiving treatment, 315 of 341 patients (92.4%) in the insulin detemir
group and 156 of 164 patients (95.1%) in the NPH insulin group com- pleted the trial. A similar
proportion of patients with- drew from each treatment group due to ineffective therapy, adverse events
and non-compliance with the protocol. The overall nature and proportion of different pretrial insulin
regimens was similar for the two treat- ment groups. At the end of the treatment period, the percentage
of patients receiving once or twice-daily basal insulin injections was 39 and 61%, respectively, for the
insulin detemir group, and 41% and 59%, respectively, for the NPH insulin group. Secara
keseluruhan, dua kelompok terapi terdistribusi merata (seimbang) secara demografi dan
karakteristik/ nilai baseline (nilai pengukuran sebelum dilakukan penelitian). 315 dari 341 pasien
(92.4%) pada kelompok inslulin detemir dan 156 dari 164 pasien (95.1%) pada kelompok insulin
NPH mengikuti penelitian secara lengkap (tidak drop out sampai final). Proporsi yang seimbang
antara 2 kelompok disebabkan oleh tidak efektifnya terapi yang diterima, kejadian efek samping
obat, serta ketidakpatuhan terhadap protocol penelitian. Regimen insulin sebelum dilakukan
penelitian pada dua kelompok juga hampir sama. Pada akhir periode terapi persentase pasien
yang menerima 1 dan 2 kali sehari insulin basal adalah 39% dan 61% untuk kelompok insulin
detemir serta 41% dan 59% untuk kelompok insulin NPH.

Glycaemic Control
Significant reductions from baseline in HbA1c levels were observed for insulin detemir (0.2%-points;
p¼0.004) and NPH insulin (0.4%-points; p¼0.0001) following 26weeks of treatment, with mean
HbA1c con- centrations comparable for the two basal insulin regi- mens at the end of the trial (insulin
detemir, 7.6%; NPH insulin, 7.5%) (table2). The number of basal insu- lin injections administered per
day (1 or 2) did not sig- nificantly affect HbA1c levels in either treatment group (p¼0.50). After
26weeks of treatment, mean FPG levels were significantly reduced from baseline by insulin detemir (by
0.5mmol/l; p¼0.027) and NPH insulin (by 0.6mmol/l; p¼0.026); mean FPG concentrations did not
differ significantly between the two treatment groups at the end of the treatment period (p¼0.66)
(table2). Moreover, no significant difference was observed in the overall shapes of the 9-point SMBG
profiles for the two treatments, measured after 26weeks of treatment (p¼0.58) (figure1). By contrast,
within- subject day-to-day variation in fasting blood glucose concentrations, based on measurements
(SMBG) recorded during the final week of the 26-week treatment period, was significantly lower for
insulin detemir than for NPH insulin (SD: insulin detemir, 1.3mmol/l; NPH insulin, 1.4mmol/l;
p¼0.021) (table3). Penurunan kadar HbA1c secara signifikan dari nilai baseline setelah 26
minggu terapi diobservasi pada kelompok insulin detemir (0,2% point; p=0,004) dan insulin NPH
(0,4% point; p=0,0001). Selanjutnya, nilai rata-rata HbA1c pada dua kelompok dibandingkan di
akhir terapi (insulin detemir 7,6%; insulin NPH 7,5%). Frekuensi pemberian injeksi insulin basal
per hari (1 atau 2) tidak mempengaruhi kadar HbA1c secara signifikan pada 2 kelompok
(p=0,50). Setelah 26 minggu terapi, kadar rata-rata FPG secara signifikan menurun dari nilai
baseline dengan insulin detemir (0.5mmol/l; p=0.027) dan insulin NPH (0.6 mmol/l; p=0.58).
Sebaliknya, variasi FPG pada subjek dari hari ke hari (berdasarkan pengukuran SBMG selama
minggu terakhir pada periode 26 minggu terapi) secara signifikan lebih rendah pada kelompok
insulin detemir dibandingkan dengan kelompok insulin NPH (SD: insulin detemir, 1.3mmol/l;
NPH insulin, 1.4mmol/l; p=0.021)

Insulin Doses
Mean doses of basal and total insulin during the 26-week treatment period are shown in figure2. After
26weeks of treatment, mean daily doses of insulin detemir and NPH insulin were 36.4U and 35.3IU,
respectively. Mean daily doses of bolus insulin (insulin aspart) after 26weeks of treatment were 40.2U
for the insulin detemir group and 35.8U for NPH insulin group. Rata-rata dosis basal dari dosis total
insulin dapat dilihat di gambar 2. Setelah 26 minggu terapi, rata-rata dosis harian insulin detemir
dan insulin NPH adalah 36.4U dan 35.3IU. Rata-rata dosis harian insulin bolus (insulin aspart)
setalah 26 minggu terapi adalah 40.2U untuk kelompok insulin detemir dan 35.8U untuk
kelompok insulin NPH.

Hypoglycaemic Episodes
The percentage of patients who experienced hypogly- caemic episodes during the final 22weeks of
treatment was comparable for the two treatment groups (table4). Very few patients (<2%) in either
treatment group experienced a major hypoglycaemic episode. The rela- tive risks of experiencing
hypoglycaemic episodes over a 24-h period or a nocturnal hypoglycaemic episode were also similar for
the two treatments (table4). The relative risk of experiencing a major hypoglycaemic episode was not
calculated in the present study due to a very low number of episodes. Persentase pasien yang
mengalami hipoglikemia selama 22 minggu terapi dibandingkan antara 2 kelompok subjek
penelitian. Sangat sedikit pasien (<2%) pada dua kelompok subjek penelitian yang mengalami
hipoglikemia major. Nilai relative risk kejadian hipoglikemia selama 24 jam atau nocturna
hipoglikemia juga hampir sama pada dua kelompok subjek penelitian. Relative risk (risiko
relative) kejadian hipoglikemia major tidak dicantumkan dalam penelitian ini karena nilainya
sangat kecil.

Body Weight
Patients treated with insulin detemir gained signifi- cantly less weight during the 26-week treatment
period compared with those receiving NPH insulin [mean difference in body weight (adjusted for body
weight at baseline):0.79kg (1.44; 0.14),p¼0.017,see figure3]. This difference in weight gain was not
due to individual changes in HbA1c levels; with adjustment for change from baseline in HbA1c, the
mean difference in body weight (insulin detemirNPH insulin) was 0.77kg (1.41; 0.12) (p¼0.020).
Pasien yang diterapi dengan insulin detemir mengalami penurunan berat badan secara signifikan
selama 26 minggu terapi dibandingkan dengan kelompok terapi insulin NPH (perbedaan rata-
rata berat badan (dibandingkan dengan berat badan baseline):≥0.79kg (≥1.44; ≥0.14),p=0.017,
lihat gambar 3). Perbedaan ini tidak bergantung pada kadar perubahan HbA1c. Dengan
perubahan nilai baseline HbA1c, perbedaan rata-rata berat badan (insulin detemir-NPH insulin)
adalah -0.77kg (-1.41; -0.12) (p=0.020).

Evaluation of Safety
Insulin detemir and NPH insulin were generally well tolerated, with the overall incidence, pattern and
severity of adverse events similar for the two treatment groups. For both treatments, the majority
(>90%) of adverse events were of mild or moderate severity. Only a small proportion of adverse events
(5% of patients on insulin detemir and 3% of patients on NPH insulin) were considered possibly or
probably related to the trial product. Of these, gastro-intestinal disorders (pain, nausea and vomiting)
were most common in patients receiving insulin detemir (1.5% of patients), whereas skin and
appendage disorders (pruritus, rash erythem- atous and urticaria) were most frequent in patients
receiving NPH insulin (1.8% of patients). Only one serious adverse event (hypoglycaemia),
experienced by a subject receiving NPH insulin, was considered related to trial product. One death
occurred during the trial (a subject in the insulin detemir group with a history of coronary heart
disease), but this was not considered related to the trial product. No clinically relevant findings in
clinical laboratory tests, physical examinations, vital signs, fundoscopy/ fundusphotograpy or 12-lead
ECG examinations were observed for either treatment group. Insulin detemir dan insulin NPH secara
umum dapat dtoleransi dengan baik, dengan kejadian efek samping yang hampir sama pada dua
kelompok subjek penelitian. Untuk dua kelompok treatment sebagian besar efek samping yang
terjadi adalah efek samping ringan sampai sedang (>90%). Hanya sedikit kejadian efek samping
(5% pasien pada kelompok insulin detemir dan 3% pasien pada kelompok insulin NPH) yang
mengalami kemungkinan mengalami efek samping akibat produk penelitian. Gangguan
gastrointestinal (nyeri abdomen, mual, muntah) adalah efek samping yang paling umum terjadi
pada pasien yang mendapatkan terapi insulin detemir (1.5%). Sedangkan gangguan kulit
(pruritus, kemerahan dan urtikaria) paling sering terjadi pada pasien yang menerima insulin
NPH (1.8%). Hanya 1 kejadian hipoglikemia serius yang terjadi pada subjek dengan terapi
insulin NPH yang diduga berhubungan dengan produk penelitian. 1 kasus kematian terjadi
selama penelitian berlangsung (subjek pada kelompok insulin detemir dengan riwayat penyakit
jantung koroner), namun hal ini tidak berhubungan dengan produk penelitian. Tidak ada
hubungan klinis yang ditemukan pada dua kelompok subjek penelitian terkait hasil pemeriksaan
klinis laboratorium, pemeriksaan fisik, tanda-tanda vital, fundoskopi/ fundusphotography atau
pemeriksaan 12-lead ECG.

Discussion
The present study in patients with type 2 diabetes com- pared the efficacy and safety of a basal-bolus
insulin regimen comprising either insulin detemir or NPH insulin in combination with mealtime insulin
aspart. Following 26weeks of treatment, comparable glycaemic control was observed for the two
treatment regimens; insulin detemir and NPH insulin reduced HbA1c and FPG levels by a similar
extent and had closely matched self-measured 9-point blood glucose profiles.
Consistent with studies in patients with type 1 diabetes, a notable difference between insulin detemir
and NPH insulin was the significantly lower within-subject day-to-day variation in fasting blood glu-
cose levels associated with insulin detemir. Glycaemic variability could arise from erratic compliance to
insulin treatment or could, for example, be due to recurrent lapses of coexisting illnesses or drug–drug
interactions. However, the reduced within subject variability associated with insulin detemir is thought
to be due to a more consistent release of insulin from the subcutaneous depot because, in contrast to
NPH insulin, insulin detemir remains in solution [13]. This characteristic also circumvents the
inconsistency in administered dose that arises from inadequate resuspension of insulin prepar- ations
(e.g. NPH insulin) prior to injection [11]. To our knowledge, insulin detemir is the first basal insulin to
demonstrate a within-subject variability in fasting blood glucose that is significantly lower than NPH
insulin in patients with type 2 diabetes. This attribute may be of particular benefit in the treatment of
type 2 diabetes because long-term glucose instability, rather than the severity of hyperglycaemia or
progression to high or low FPG levels over time, is a predictor of cardiovascu- lar-related mortality in
elderly (>75years) patients [28]. Furthermore, the lower day-to-day variation and hence more
predictable glycaemic response to insulin detemir should increase confidence in titrating insulin doses
to achieve more ambitious blood glucose targets due to a reduced fear of hypoglycaemia [29]. The
lower within-subject variability associated with insulin detemir is anticipated to reduce the frequency
by which blood glucose levels fall into the hypoglycaemic range (<2.8mmol/l). In keeping with this, a
lower proportion of patients experienced episodes of hypoglycaemia in the insulin detemir group.
However, the risks of overall and nocturnal hypoglycaemia were not significantly different for insulin
detemir and NPH insulin. This may be related to the generally lower overall incidence of
hypoglycaemia in type 2 diabetes (com- pared to type 1 diabetes [30]) and the fact that the study was
powered to detect differences between treat- ment groups in HbA1c rather than hypoglycaemia. Despite
comparable levels of glycaemic control, patients administering insulin detemir gained signifi- cantly
less body weight than those receiving NPH insu- lin, corroborating previous findings for patients with
type 1 diabetes [17–19,21]. Fokus penelitian ini adalah membandingkan efektifitas dan keamanan
regimen insulin basal bolus antara insulin detemir dan insulin NPH yang dikombinasikan dengan
insulin aspart yang diberikan pada saat makan pada pasien DM tipe 2. Selama 26 minggu terapi,
kadar glukosa darah dibandingkan antara 2 kelompok tersebut. Penurunan kadar HbA1c dan
FPG serta profil 9 point SBMG hampir sama pada 2 kelompok. Sejalan dengan penelitian pada
pasien DM tipe 1, insulin detemir memberikan penurunan secara signifikan variasi subjek pada
pengukuran FPG dibandingkan dengan insulin NPH. Variablilitas kadar glukosa darah
bergantung dapat meningkat bergantung pada kepatuhan pasien terhadap regimen insulin yang
diberikan, adanya penyakit comorbid lainnya atau terjadinya interaksi dengan obat lain. Namun
demikian, penurunan variabilitas subjek pada penggunaan insulin detemir diduga disebaban
karena pelepasan insulin yang lebih konsisten. Selain itu, berbeda dengan insulin NPH, insulin
detemir tetap dalam bentuk larutan. Hal ini menyebabkan inkonsistensi insulin NPH saat
diadministrasikan akibat resuspensi yang tidak adekuat saat penyiapan insulin untuk
diinjeksikan (e.g insulin NPH). Insulin detemir merupakan insulin basal pertama yang
didemonstrasikan bahwa insulin detemir dapat menurunkan variabilitas individu subjek pada
pengukuran FPG dibandingkan insulin NPH pada DM tipe 2. Hal ini dapat memberikan
keuntungan pada pengobatan DM tipe 2 terkait progresi peningkatan atau penurunan kadar
FPG yang menjadi predictor mortalitas yang berkaitan dengan cardiovascular event pada pasien
elderly (>75 tahun). Selain itu, hal tersebut juga dapat menjadi acuan dalam mentitrasi dosis
insulin untuk mencapai kadar glukosa darah mendekati normal tanpa khawatir mengalami
hipoglikemia. Sehingga proporsi pasien yang mengalami hipoglikemia pada kelompok insulin
detemir lenbih rendah daripada insulin NPH. Namun, risiko secara keseluruhan dan risiko
nocturnal hypoglicaemia tidak berbeda secara signifikan anatar kelompok insulin detemir dan
insulin NPH. Hal ini berhubungan dengan rendahnya insiden hipoglikemia pada DM tipe 2
dibandingkan dengan DM tipe 1. Selain itu, penelitian ini lebih menitikberatkan pada perbedaan
nilai HbA1c dibandingkan dengan risiko hipoglikemia. Pasien dengan insulin detemir juga
mengalami penurunan berat badan yang lebih signifikan dibandingkan dengan insulin NPH. Hal
ini senada dengan hasil penelitian sebelumnya pada pasien DM tipe 1. The mechanism(s) behind
this reduced weight gain are currently unknown. Because reductions in both weight gain and hypogly-
caemic risk are a regular finding of clinical studies with insulin detemir [17–19,21], we have proposed
that the lower weight gain associated with insulin detemir may result from a decreased need to
counteract hypogly- caemia through defensive, between-meal snacking, although recognize that other
yet unknown factors may also contribute. Mekanisme yang melatarbelakangi penurunan berat
badan pada pemberian insulin detemir ini masih belum diketahui. Kami menduga, penurunan
berat badan disebabkan penurunan kebutuhan untuk mengantisipasi kejadian hipoglikemia
melalui mekanisme defensive antara makan dan snack ringan meskipun factor lain yang masih
belum diketahui juga kemungkinan memiliki kontribusi tersendiri. In the current study, the risk of
hypogly- caemia was not significantly different between the two treatment groups. However, a lower
proportion of patients experienced hypoglycaemic episodes in the insulin detemir group. It could
therefore be speculated that the reduced weight gain observed for insulin detemir resulted, at least in
part, from a reduction in defensive, between-meal snacking. However, further studies are required to
confirm or refute this hypothesis. Pada penelitian ini, risiko terjadinya hipoglikemia pada 2
kelompok tidak berbeda secara signifikan. Namun, proporsi pasien yang mengalami
hipoglikemia pada kelompok insulin detemir lebih kecil. Weight gain can be a major drawback of
long-term intensive insulin therapy [24] as the accumulation of visceral fat (central obesity), for
example, is closely asso- ciated with insulin resistance, hypertension and dys- lipidaemia [31–33].
Weight-induced changes in lipid profile and blood pressure increase the risk of coronary artery disease
[34,35]. Moreover, studies indicate a higher prevalence of a number of complications asso- ciated with
type 2 diabetes in obese vs. non-obese patients [36]. Peningkatan berat badan dapat menjadi
manifestasi major penggunaan insulin jangka panjang akibat terjadinya akumulasi lemak
visceral (central obesity). Hal ini juga berhubungan dengan mekanisme resistensi reseptor
insulin, hipertensi dan dislipidemia. Peningkatan berat badan berhubungan dengan peningkatan
kadar lipid sehingga dapat meningkatkan risiko cardiovascular disease. Accordingly, a basal
insulin, such as insulin detemir, that is associated with less weight gain could offer an important
advantage in the management of type 2 diabetes. In summary, patients with type 2 diabetes receiving 26
weeks of once- or twice-daily basal injections with insu- lin detemir plus insulin aspart at mealtimes
experienced comparable glycaemic control but significantly lower within-subject variability in fasting
blood glucose and less weight gain compared to patients treated with NPH insulin and insulin aspart.
Insulin detemir was well tolerated and had a similar safety profile to NPH insulin. Thus, insulin detemir
represents a promising new basal insulin that could offer distinct advantages over NPH insulin in the
treatment of patients with type 2 diabetes.
 Introduction
leads to long-term microvascular and macrovascular
Type 2 diabetes is a chronic metabolic disorder, complications that substantially increase morbidity
currently and
affecting 140–150 million people worldwide, mortality [1].
which often
Correspondence:
Prof Thomas Haak, Research Institute of the Diabetes Academy Mergentheim, Theodor-Klotzbu¨cher-
Str. 12, D-97980 Bad
Mergentheim, Germany.
E-mail:
haak@diabetes-zentrum.de
56
Diabetes, Obesity and Metabolism, 7, 2005, 56–64 # 2004 Blackwell
Publishing Ltd
T. Haak et al. plasma Insu
concentration approximately 5 h after lin
injection and has
Traditionally, oral antidiabetic drugs (OADs) an insufficient duration of action to cover dete
consti- night-time mir
tute the first line of treatment for type 2 diabetes requirements [7,8]. Furthermore,
when in
inadequate resus-
diet and exercise fail to provide satisfactory pension prior to injection and variations in patie
glycaemic crystal size nts
control. However, as the disease advances, make dosing precision and absorption
progressive with
kinetics highly
loss of pancreatic b-cell function necessitates the variable, resulting in unpredictable glucose type
use of levels [9–11]. 2
insulin therapy when glycaemic targets cannot These limitations have therefore diab
be prompted the devel-
achieved by oral therapy alone [2]. Although etes
opment of new basal insulin analogues
insulin (insulin glargine
has commonly been regarded as a final treatment and insulin detemir [12]) of which insulin properties
option, detemir, a that are
several recent studies have provided evidence derivative of human insulin soluble at associated
that tight with lower
neutral pH within-subject
glycaemic control, achieved in part through [13,14], is the most recent to be developed variability in
earlier and [15]. The fasting blood
more intensive use of insulin, can reduce the protracted action of insulin detemir largely glucose and a
incidence stems from reduced risk
and delay the progression of a number of long- delayed absorption from the injection site of
term due to self- hypoglycaemi
complications associated with type 2 diabetes association and binding, via a covalently a [17–21].
[3–5]. attached fatty Furthermore,
Moreover, evidence suggests that in contrast to acid moiety, to albumin in the interstitial compared to
sulphon- fluid [13,14]. NPH
ylureas, early insulin treatment in newly This novel mechanism of protraction insulin,
diagnosed provides more patients
patients with type 2 diabetes may temporarily reproducible insulin absorption and lower receiving
prolong within-sub- insulin
endogenous insulin secretion and promote better ject variability than other basal insulins detemir gain
meta- [16]. signifi-
bolic control [6]. Comparative studies of insulin detemir cantly less
Usually, insulin therapy is initiated as a daily and NPH insu- body weight
injec- lin in patients with type 1 diabetes have at comparable
tion of long-acting basal insulin in combination demonstrated metabolic
with similar glycaemic control but a more control
OADs. However, when glycaemic control extended and [17–
becomes smoother time-action profile with insulin 19,21,22].
inadequate, a basal-bolus insulin regimen is detemir, These
often characteristics
adopted [2]. Although the development of short- could be
acting # 2004 Blackwell Publishing Ltd
particu-
bolus insulin analogues (e.g. insulin aspart and larly
insulin advantageous
lispro) has improved postprandial glycaemic in the
control, treatment of
traditional basal insulin preparations [e.g. neural type 2
prot- diabetes
amine hagedorn (NPH) insulin and where
Ultralente1insulin] approximatel
cannot deliver insulin at the constant and y 80% of
reproducible newly
low level that characterizes normal insulin diagnosed
secretion. patients
For example, NPH insulin reaches a peak in are
overweight or obese [23]. For patients with type 2 detemir (LevemirTM, Novo Nordisk A/S; 100 OA
diabetes, fears of weight gain and hypoglycaemia U/ml) or
asso- NPH insulin (isophane human insulin, Novo
ciated with intensive insulin therapy [24] have often Nordisk
resulted in a reluctance to initiate insulin treatment
and/or titrate insulin to a dose where near-normal Diabetes, Obesity and Metabolism, 7,
gly- 2005, 56–64
caemia is achieved [25].
The present study was conducted to compare the
effi-
cacy, safety and weight change associated with a
basal-
bolus regimen comprising either insulin detemir or
NPH
insulin in combination with mealtime insulin aspart
in
patients with type 2 diabetes previously treated with
insulin.

Research Design and Methods

Sixty-three sites in five European countries
participated
in this open-label, parallel-group trial. The trial con-
sisted of a 3-week screening period and a 26-week
treat-
ment period. The study was approved by local ethics
committees and health authorities according to local
regulations and was conducted in accordance with
ICH-Good Clinical Practice guidelines [26] and the
Declaration of Helsinki [27]. Signed informed
consent
was provided before any trial-related activities.

Patients
A total of 505 patients with type 2 diabetes of
12 months
duration, aged 35 years, HbA1c 12.0% and who
had received insulin treatment for at least 2 months
(basal insulin dose 30% of the total daily insulin
dose)
were included in the trial. Patients were excluded if
they
had received OADs within 2 months of the trial,
were
pregnant or breast feeding, suffered from
proliferative
retinopathy, uncontrolled hypertension, recurrent
major
hypoglycaemia, impaired renal or hepatic function,
cardiac problems, or if they received a total daily
basal
insulin dose > 100 IU/day.

Study Design 57
Patients were randomized (2 : 1) to receive either
insulin
OA Insulin detemir in patients with type 2 deter-
diabetes mined in whole blood by high-performance liquid
chro-
matography (HPLC) using a Bio-Rad
A/S, Copenhagen, Denmark; 100 IU/ml) as basal Variant TM
insulin
for 26 weeks. Patients receiving more than one instrument. Clinic fasting plasma glucose (FPG)
basal concen-
insulin injection per day prior to the trial trations were determined by a hexokinase method.
received These and standard safety determinations
twice-daily (before breakfast and at bedtime) (haematology
injections and biochemistry) were performed by a central labora-
following randomization. All other patients, tory (Laboratorium fu¨r Klinische Forschung GmbH,
including Raisdorf, Germany). One Touch Profile1blood glucose
those receiving biphasic (premixed insulin) prior meters (LifeScan, Neckargemu¨nd, Germany)
to the calibrated
trial, received a once-daily basal insulin for blood glucose measurements were used for all
injection (at SMBG assessments. HbA1c and FPG were measured
bedtime). For all patients, insulin aspart 1 at
, the screening visit and after 13 and 26 weeks of treat-
(NovoRapid
Novo Nordisk A/S, 100 U/ml) was administered ment. Patients performed 9-point SMBG profiles
imme- before
diately prior to each main meal. All insulin randomization and after 13 and 26 weeks of treatment.
preparations In
were injected subcutaneously (basal insulin, addition, fasting SMBG levels were measured on each
abdomen of
or thigh; bolus insulin, abdomen) using a the final 7 days of treatment.
Adverse events, coded according to the Novo
NovoPen13 Nordisk
device (Novo Nordisk A/S). Throughout the Adverse Reaction Dictionary, were monitored
trial, basal through-
insulin doses were optimized according to self- out the study period and, where necessary, were
mea-
sured blood glucose (SMBG) values. Patients treated by established methods of care. Treatment emergent
rando-
mized to insulin detemir were initiated on 50% adverse events were defined as events occurring from
of their
pretrial basal insulin dose whereas those Diabetes, Obesity and Metabolism, 7, 2005, 56–64
randomized to
NPH insulin continued with their pretrial basal
insulin
dose. During the trial, patients were instructed to
aim for
the following SMBG levels: prebreakfast, 4–7
mmol/l;
nocturnal (02 : 00–04 : 00), 4–7 mmol/l;
postprandial
(90 min after meal), <10 mmol/l. If a patient
did not
reach these SMBG targets on a once-daily basal
insulin
regimen (either with insulin detemir or NPH
insulin),
and further dose escalations were inadvisable,
they were
transferred to a twice-daily basal insulin regimen
with
58 appropriate dose adjustments.

Methods
HbA1c (reference range of assay: 4.3–6.1%) was
 f, as
T. Haak et al. following 26 weeks of treatment were analysed using a
minor if blood glucose concentration was <2.8 similar ANOVA model to that used for the
mmol/l primary end-
the first but no assistance was required, and as symptoms point. The 9-point SMBG profiles, measured after
day of 26 weeks of treatment, were analysed using a repeated
dosing to 7 only if
days the episode was not confirmed by a blood measures ANOVA model. Within-subject day-to-day
following glucose meas- varia-
the final tion in fasting SMBG levels during the last 7 days of
dose. urement. Nocturnal hypoglycaemia was defined
Clinical as any treatment was compared between the two treatment
laboratory hypoglycaemic episode that occurred between groups using variance component models. To estimate
assessment the hours the relative risk of hypoglycaemia, all hypoglycaemic
s of 23 : 00 and 06 : 00.
(haematolo # 2004 Blackwell
gy and Statistical Analyses Publishing Ltd
bio-
chemistry) All analyses were based on the intention-to-treat
, physical (ITT)
examinatio analysis set, which included all patients
ns, randomized
fundoscop and exposed to at least one dose of study drug.
y/fundus The
photograp primary endpoint, HbA1c (%) after 26 weeks
hy, 12-lead of treat-
electrocard ment, was evaluated by an analysis of variance
iogram (ANOVA)
(ECG), model, with treatment and country as fixed
and meas- effects and
urements covariate adjustment for baseline values. The
of vital study had
signs sufficient power (85%) to detect an average
(blood difference of
pressure 0.4% in HbA1c between treatment groups.
and pulse) A 95% two-
and sided confidence interval (CI) was constructed
body for the
weight difference between the treatment group means
were (insulin
performed detemir NPH insulin); insulin detemir was
at deemed
screening, non-inferior if the upper limit of the 95% CI was
randomiza <0.4% (absolute). Treatments were considered
- compar-
tion and able if the non-inferiority criterion was fulfilled.
after 13 The influence of the number of basal insulin
and 26 injec-
weeks of tions per day on HbA1c levels following
treatment 26 weeks of
(as appro- treatment was investigated by including the
priate). number of
Hypogly basal injections as a factor in the ANOVA
caemic model used for
episodes the primary endpoint while successively
were including and
classifiedexcluding the treatment by number of basal
as major injections
when a interaction term.
subject Change in body weight (both with and without
was unable adjust-
to treat ment for change in HbA1c concentration) and
him/hersel FPG levels
T. Haak et al. Insulin detemir in patients with type 2 OA
diabetes
episodes occurring during the maintenance
period were
analysed as recurrent events using a gamma p ¼ 0.004) and NPH insulin (0.4%-points; p ¼
0.0001)
frailty following 26 weeks of treatment, with mean
model with treatment group as covariate. HbA1ccon-
Nocturnal centrations comparable for the two basal
hypoglycaemic episodes were analysed insulin regi-
separately mens at the end of the trial (insulin detemir,
(as above). All results are presented as means 7.6%;
SD NPH insulin, 7.5%) (table 2). The number of
unless otherwise specified. The significance basal insu-
level was 5%. lin injections administered per day (1 or 2) did
not sig-
Results nificantly affect HbA1c levels in either
treatment group
Overall, the two treatment groups were evenly (p ¼ 0.50).
matched After 26 weeks of treatment, mean FPG
with respect to demographic and baseline levels were
characteristics significantly reduced from baseline by insulin
(table 1). Of those receiving treatment, 315 of detemir
341 (by 0.5 mmol/l; p ¼ 0.027) and NPH insulin
patients (92.4%) in the insulin detemir group and(by
156 0.6 mmol/l; p ¼ 0.026); mean FPG
of 164 patients (95.1%) in the NPH insulin concentrations did
group com- not differ significantly between the two
pleted the trial. A similar proportion of patients treatment
with- groups at the end of the treatment period (p ¼
drew from each treatment group due to 0.66)
ineffective (table 2). Moreover, no significant difference
therapy, adverse events and non-compliance was
with the observed in the overall shapes of the 9-point
protocol. The overall nature and proportion of SMBG
different profiles for the two treatments, measured after
pretrial insulin regimens was similar for the two 26 weeks
treat- of treatment (p ¼ 0.58) (figure 1). By contrast,
ment groups. At the end of the treatment period, within-
the subject day-to-day variation in fasting blood
percentage of patients receiving once or twice- glucose
daily concentrations, based on
basal insulin injections was 39 and 61%, measurements (SMBG)
respectively, recorded during the final week of the 26-week
for the insulin detemir group, and 41% and 59%, treatment
respectively, for the NPH insulin group. period, was significantly lower for insulin
detemir than
for NPH insulin (SD: insulin detemir, 1.3
Glycaemic Control mmol/l; NPH
Significant reductions from baseline in HbA1c insulin, 1.4 mmol/l; p ¼ 0.021) (table 3).
levels
were observed for insulin detemir Insulin Doses
(0.2%-points; Mean doses of basal and total insulin during
the 26-week
treatment period are shown in figure 2. After
26 weeks of

Table 1 Demographics and baseline characteristics of patients with type 2 diabetes exposed to trial medication
Number of patients
exposed
Ethnic origin Insulin detemir NPH insulin
Caucasian 341 164
Asian-Pacific islander
Gender (male/female) 338 162
Age (years) 3 2
Weight (kg)2 165/176 93/71
BMI (kg/m ) 60.68.7 60.08.4
Duration of type 2 diabetes (years) 85.714.9 89.317.5
HbA1c(%) 30.15.0 31.15.8
FPG (mmol/l) 12.97.4 13.78.0
Pretrial insulin regimen (%) 7.9 1.3 7.8 1.3
Basal-bolus 10.13.32 10.43.42
Biphasic (premixed)
Mean daily pretrial insulin dosez 86 88
Basal 14 12
Bolus
27.814.7* 28.015.4y
33.619.2* 34.819.9*
Values are absolute numbers, percentages or means SD. NPH, neural protamine hagedorn.
*Units (U).
yInternational units (IU).
zDoes not include patients treated with premixed insulin formulations.

# 2004 Blackwell Publishing Ltd Diabetes, Obesity and Metabolism, 7, 59

2005, 56–64
OA Insulin detemir in patients with type 2 T. Haak et al.
diabetes

Table 2 Glycaemic control after 26 weeks of Insulin detemir NPH

treatment insulin, mean difference Significan
[95% CI] ce
Insulin detemir NPH insulin
HbA1c(%)7.6 0.1 (n ¼315) 7.5 0.1 (n ¼ 155)0.16 [0.003; 0.312] Insulin detemir non-inferior*
FPG (mM) 9.7 0.2 (n ¼309) 9.6 0.3 (n ¼ 152)0.11 [ 0.400; 0.630] p ¼ 0.66y
Values are estimated (least square) means SE. NPH, neural protamine hagedorn.
*Non-inferiority criterion: upper limit of 95% CI < 0.4% HbA1c.
yANOVA with treatment and country as fixed effects and covariate adjustment for baseline.

treatment, mean daily doses of insulin detemir 9

and NPH
insulin were 36.4 U and 35.3 IU, respectively. 8
Mean
daily doses of bolus insulin (insulin aspart) after 7
26 weeks of treatment were 40.2 U for the
insulin 6
detemir group and 35.8 U for NPH insulin
group. Fig. 1 Mean
9-point blood glucose pro- 5
Hypoglycaemic Episodes
The percentage of patients who experienced
hypogly-
caemic episodes during the final 22 weeks of
treatment
was comparable for the two treatment groups
(table 4).
Very few patients (<2%) in either treatment
group
experienced a major hypoglycaemic episode.
The rela-
tive risks of experiencing hypoglycaemic
episodes over a
24-h period or a nocturnal hypoglycaemic
episode were
also similar for the two treatments (table 4). The
relative
risk of experiencing a major hypoglycaemic
episode was
not calculated in the present study due to a very
low
number of episodes.

Body Weight
Patients treated with insulin detemir gained
signifi-
cantly less weight during the 26-week treatment
period
compared with those receiving NPH insulin
[mean

10
 groups. For both treatments, the majority (>90%) of
difference in body weight (adjusted for body weight at adverse events were of mild or moderate severity.
baseline): 0.79 kg ( 1.44; 0.14), p ¼ 0.017, see Only
figure 3]. a small proportion of adverse events (5% of patients
This difference in weight gain was not due to individual on
changes in HbA1c levels; with adjustment for change insulin detemir and 3% of patients on NPH insulin)
from baseline in HbA1c, the mean difference in body were considered possibly or probably related to the
weight (insulin detemir NPH insulin) was trial product. Of these, gastro-intestinal disorders
0.77 kg (pain,
( 1.41; 0.12) (p ¼ 0.020). nausea and vomiting) were most common in patients
receiving insulin detemir (1.5% of patients), whereas
skin and appendage disorders (pruritus, rash erythem-
Evaluation of Safety atous and urticaria) were most frequent in patients
Insulin detemir and NPH insulin were generally well receiving NPH insulin (1.8% of patients). Only one
tolerated, with the overall incidence, pattern and sever- serious adverse event (hypoglycaemia), experienced
ity of adverse events similar for the two treatment by
files measured during the final week 08:00 10:00 12:00 14:00 16:00 18:00 20:00 22:00
of 24:00 02:00 04:00
treatment with insulin detemir (&) or
neural protamine hagedorn (NPH) Breakfa
st
Lunch Dinner Bedtime
insulin (&). Bars indicate SEM. Time

60 Diabetes, Obesity and # 2004 Blackwell

Metabolism, 7, 2005, 56–64 Publishing Ltd
Blood glucose (mM)