O

CdT

KVa

b

a

A

R

R

2

A

A

K

D

T

I

C

R

1

Ttmb

1h

p r i m a r y c a r e d i a b e t e s 7 ( 2 0 1 3 ) 235–242

Contents lists available at ScienceDirect

Primary Care Diabetes

j o u r n a l h o m e p a g e : h t t p : / / w w w . e l s e v i e r . c o m / l o c a t e / p c d

riginal research

linical outcomes and costs for patients with type 2iabetes mellitus initiating insulin therapy in Greece:wo-year experience from the INSTIGATE study

yriakos Aloumanisa,∗,1, Marianna Benroubib,1, Sotiria Sourmelia,1,angelis Drossinosa,1

European Medical Research Institute by Pharmaserve-Lilly, Athens, GreeceDiabetes Department, Polyclinic General Hospital, Athens, Greece

r t i c l e i n f o

rticle history:

eceived 10 October 2012

eceived in revised form

5 February 2013

ccepted 1 April 2013

vailable online 25 April 2013

eywords:

iabetes

reatment

nsulin

ost

a b s t r a c t

Aims: To evaluate the quality of metabolic control, clinical outcomes, resource costs, and

quality of life among patients with type 2 diabetes mellitus (T2DM), who initiated insulin

for the first time as part of routine clinical practice.

Methods: The INSTIGATE study is a prospective, multicentric, observational study of patients

initiating insulin treatment. This sub-cohort analysis focuses on Hellenic outcomes.

Results: At baseline, 263 Greek patients were enrolled just before initiating insulin for the

first time. At the 6-month visit, 237 patients (90.1%) remained and consented to an addi-

tional 18-month observation period. In these 237 extension patients, over the 24-month

post-initiation period, HbA1c (mean(SD)) decreased from 9.7%(1.6%) to 7.1%(0.9%) and body

weight and BMI increased (+3(6) kg and +1.1(2.2) kg/m2, respectively). At each post-baseline

visit approximately one in five patients reported ≥1 episodes of hypoglycaemia in the pre-

ceding 3–6 months. Median total costs fluctuated from 438D at baseline to 538D up to 6

esources months and 451D at 24 months; mean costs were 496(383)D, 573(276)D and 485(247)D, respec-

tively.

Conclusions: In this cohort, insulin treatment seems to be effective with little long-term

impact on cost. Findings should be interpreted in the context of an observational study.

ry Ca

by 2030 [2], and causes a major impact on personal well-being,

© 2013 Prima

. Introduction

ype 2 diabetes mellitus (T2DM) is a chronic metabolic disease

hat, when poorly controlled, can lead to serious micro- and

acro-vascular complications affecting many organs [1]. Dia-etes imposes a large burden on the individual patient, the

∗ Corresponding author at: 15th Klm National Road Athens-Lamia, 1456E-mail addresses: aloumanisk@hotmail.com, aloumanis kyriakos@l

1 For the Hellenic INSTIGATE study group.751-9918/$ – see front matter © 2013 Primary Care Diabetes Europe. Puttp://dx.doi.org/10.1016/j.pcd.2013.04.001

re Diabetes Europe. Published by Elsevier Ltd. All rights reserved.

national healthcare systems and the states’ economies, withan estimated world prevalence of 285 million in 2010 (6.6% ofthe world’s adult population), predicted to rise to 438 million

4 Kifissia, Greece. Tel.: +30 2106294618; fax: +30 2106294850.illy.com (K. Aloumanis).

social function and treatment costs [3–5].Better glycaemic control has repeatedly been shown to be

associated with long-term reduction in risk of micro- and

blished by Elsevier Ltd. All rights reserved.

e t e s

236 p r i m a r y c a r e d i a b

macrovascular disease [1,6,7]. However, aggressive control ofhyperglycaemia in the later stages of T2DM does not appearto be associated with improved cardiovascular outcomes[8,9].

In patients with type 2 diabetes, beta pancreatic cellsshow reduced glucose sensitivity, and a reduced insulin secre-tion capacity to compensate for associated insulin resistance.Where hyperglycaemia persists over time, the glycosylation ofproteins by excess blood glucose and amyloid deposits withinthe beta cells leads to a progressive decrease in beta-cell massand further beta-cell dysfunction, exacerbating the reductionsin insulin secretion [10]. Lifestyle interventions for weight lossand physical activity followed by oral medication may not con-strain the deterioration of beta cell function [7,11], and thusmost patients will eventually require insulin therapy to main-tain glycaemic control and reduce the development of diabeticcardiometabolic complications [11]. The 2012 ADA-EASD con-sensus statement recognises the need for individualisationregarding diabetes management yet suggests that ultimatelymany patients will require insulin therapy even as a secondstep after metformin [12].

However, information regarding insulin initiation in theusual health care setting in Europe is limited [13]. Of partic-ular concern is the dearth of available information regardinghealthcare resource utilisation and cost prior to and follow-ing the beginning of insulin treatment. INSTIGATE (INSulinTItration – GAining an understanding of the burden of Type2 diabetes in Europe) study, aimed to assess the quality ofmetabolic control, clinical outcomes, and health-related qual-ity of life in patients with T2DM who were initiating insulintreatment for the first time in the course of usual medical care.Clinical outcomes and direct costs of diabetes care from Greekparticipants who were observed up to 24 months followinginsulin initiation are described.

2. Methods

2.1. Study design

INSTIGATE, a prospective, observational, non-interventional,multi-centre study, was designed to enrol patients with type2 diabetes who were initiating insulin for the first time inthe course of normal medical care, in five European countries(Germany, Greece, France, Spain, and UK). Objectives of thestudy were to assess the direct costs and to describe theresource utilisation, quality of metabolic control and clinicaloutcomes associated with the first 6 months of insulin ther-apy in patients with type 2 diabetes, with data being collectedat baseline (insulin initiation) and then after approximately 3and 6 months (published elsewhere) [14–16]. The study wassubsequently extended to include a total follow-up period of2 years in Germany, Greece and Spain following initial insulintreatment, with further data collection at approximately 12,18 and 24 months.

In Greece, 22 investigators who treated patients with dia-

betes, including diabetologists, endocrinologists, and generalpractitioners, with full responsibility for treatment changesand additional instructions regarding diabetes therapy accord-ing to their practice protocols, participated.

7 ( 2 0 1 3 ) 235–242

2.2. Study population

The study population consisted of adult patients, with T2DM,who presented within the normal course of care and who,according to the investigators’ clinical judgement, were notachieving acceptable levels of glycaemic control at the timeof enrolment, and thus were placed on insulin treatmentfor the first time. Enrollers were able to complete writtenquestionnaires and were not simultaneously participating inother similar studies. All treatment choices were made by thephysician and the patient and the insulin treatment could beco-administered with any combination of oral anti-diabetictherapy.

Local requirements for ethics were met; all patientsparticipating in the study gave written informed con-sent, according to local regulations and the Declaration ofHelsinki.

2.3. Data collected at baseline

Assessment at baseline included:

• Demographic characteristics (age, gender, ethnicity, educa-tion, socioeconomic status and smoking history).

• Diabetes-focused medical history (6-month retrospectivechart review) including age at initial diabetes diagnosis, pre-scribed medications, diet/exercise compliance, comorbidmacrovascular (coronary heart disease, previous myocar-dial infarction, chronic heart failure, peripheral arterialocclusive disease, stroke, coronary artery bypass graft, tran-sient ischaemic attack and amputation) or microvascular(diabetic retinopathy, diabetic nephropathy, diabetic neu-ropathy) conditions, and history of other diabetes-relateddiseases (such as hypertension, hyperlipidaemia, depres-sion and cancer).

• Historic (past 6 months) and current physical and labora-tory values including height, weight, body mass index (BMI),lipids, cholesterol, triglycerides, HbA1c values and fastingblood glucose.

• Historic (past 6 months) and current anti-diabetic andother related medications, including the timing and dosingstrength.

• Healthcare resource utilisation in the past 6 months (includ-ing Health Care Professional (HCP) contacts and any hospitalvisits/admissions).

• Patient reported outcomes using questionnaires to mea-sure treatment compliance (the Morisky Compliance Index),health status (EQ-5D), health-related quality of life (DHP-18)and the experience with insulin (EWIQ-I) [17–19].

• Occurrence of hypoglycaemia as reported by patients.

2.4. Statistical analysis

Statistical analyses were descriptive and exploratory in natureand included patients of the Greek subsample with at least12 months of follow-up after insulin initiation (N = 237).

Mean/standard deviation or median/interquartile range isreported for continuous variables, where appropriate.

Total direct health care costs including costs of med-ication (insulin and OADs), costs of glucose monitoring

e s 7

(tpsowa

cpcC

3

3

Ipwtop

3

Onmnltnc

3

Brwa

p r i m a r y c a r e d i a b e t

i.e. test strips), HCP visits, and costs due to hospitalisa-ions/Emergency Room (ER) visits are reported per 6-montheriods as mean and median values. For healthcare profes-ionals the median value – showing the “typical” costs isf interest while for insurance companies the mean value,hich is driven by a minority of overspending patients, is more

ppropriate.For categorical variables, absolute frequencies and per-

entages based on the number of non-missing values areresented. Missing data were not imputed. All analyses wereonducted using SAS software version 8.2 (SAS Institute Inc.,ary, NC, USA).

. Results

.1. Patient profile

n Greece 237 patients (90.1%) participated in the extensionart of the study and were included in the current analysis,ith 227 of them completing the entire 24-month observa-

ional period (86.3% of the total initial population and 95.8%f the extension population). Baseline characteristics of theseatients are shown in Table 1.

.2. Concomitant diseases

f patients initiating insulin, 15% reported no other sig-ificant medical diagnoses, while 30.4% had at least oneacrovascular and 20.7% had at least one microvascular diag-

osis. Of all enrolled patients in Greece, 72.6% reported ateast one significant diabetes-related diagnosis (e.g. hyper-ension, hyperlipidaemia) while 24.0% reported at least oneon diabetes-related significant co-morbidity (e.g. depression,ancer, etc.).

.3. Insulin initiation and regimen used

asal insulin and mixtures were the most frequently usedegimens in this subgroup: of the whole cohort 49.4% startedith long/intermediate acting insulin, 30.8% with mixtures

nd 12.2% with a basal/bolus regimen. The age, baseline HbA1c

Table 1 – Greek patients’ demographics at baseline (extension pmonths are presented) and numerical differences between grou

Number of patients 237

Mean age (SD) years 66.0(9.9)Years since diagnosis of DM (SD) 11.8(7.0)Current or previous smokers (%) 38.4No or minimum mandatory level of education (%) 75.6

N = 237 Basal insulin (117)

Age (years) 66.0(9.9)Duration of diabetes (years) 12.6(7.5)Mean(SD) most recent HbA1c

– At baseline 9.7(1.6)– At 6 months 7.6(1.2)– At 24 months 7.2(0.8)

SD, standard deviation.

( 2 0 1 3 ) 235–242 237

and diabetes duration of patients who initiated basal/bolus,basal or mixture regimens respectively are shown in Table 1.Most patients initiated insulin on an out-patient basis withonly 4.2% starting as inpatients.

Insulin treatment was intensified during the 24 monthsof follow-up, as indicated by a mean (SD) increase in totalInternational Units (IU) over time (from 31.5(16.0) at baselineto 43.1(21.1) at 24 months), although the median num-ber of administrations per day remained at 2. In the first12 months following insulin initiation, 20 patients (8.4%)switched their insulin regimen, most commonly (n = 9) fromlong/intermediate acting to a basal/bolus regimen. 9.3% ofpatients reported no use of insulin at 24 months (Fig. 1). Per-centages of patients using oral anti-diabetic medications were43.0% at insulin initiation and 52.4% at 24 months (Fig. 1).

3.4. Body weight and BMI

On average, patients were overweight at the time of insulin ini-tiation (BMI at baseline 28.2(4.7) kg/m2). At 24 months a meanincrease in body weight (+3(6) kg) and BMI (+1.1(2.2) kg/m2)was observed (Table 2). Intra-individual mean (SD) changesin BMI from baseline to 24 months by insulin regimen were+0.5(1.9) kg/m2 for long/intermediate only, +1.5(1.9) kg/m2

for mixture only, +2.4(2.9) kg/m2 for short-acting only, and+1.8(2.9) kg/m2 for basal/bolus.

3.5. Effects of treatment on HbA1c

Mean HbA1c values were 9.7%(1.6%) at baseline, 7.4%(1.1%) at6 months, and 7.1%(0.9%) at 24 months (intra-individual meanchanges of −2.2%(1.8%) and −2.5%(1.6%), respectively) (Table 2and Fig. 2). Across insulin regimens, observed mean changesin HbA1c ranged from −2.8%(1.8%) for patients on basal/bolusregimen to −2.6%(1.6%) for patients on long/intermediate act-ing and −2.5%(1.4%) on mixtures (Table 1).

3.6. Biochemistry

A mean change in fasting blood glucose (−5.5(4.0) mmol/l),LDL (−0.44(0.98) mmol/l) and Tgs (−0.45(0.87) mmol/l) was

atients followed up for >6 months with maximum 24ps with regards to chosen regimen.

Sex (% male) 52.3

BMI (SD) kg/m2 28.2(4.7)HbA1c at diagnosis (SD) % 9.5(1.8)Primary occupation (retired (%)) 44.7

Mixtures (73) Basal/Bolus (29)

67.4(9.5) 62.0(10.2)11.2(6.4) 8.6(5.8)

9.3(1.5) 10.3(1.9)7.3(0.9) 7.6(0.9)6.9(0.8) 7.2(1.1)

238 p r i m a r y c a r e d i a b e t e s 7 ( 2 0 1 3 ) 235–242

Fig. 1 – Number of ongoing/new oral anti-diabetic (OADs) medications (columns) and insulin regimens (lines) used duringstudy.

Table 2 – Insulin use, metabolic parameters and hypoglycaemic events during follow-up.

Months 0 3 6 12 18 24

Pt N 237 237 237 237 229 227HbA1c (SD) 9.7(1.6) 7.9(1.1) 7.4(1.1) 7.4(1.2) 7.3(1.0) 7.1(0.9)FBG (mmol/L) 12.8(3.9) 8.3(2.7) 7.8(2.2) 7.8(2.4) 7.5(2.1) 7.3(2.1)BMI (kg/m2) 28.2(4.7) 28.5(4.6) 28.7(4.7) 28.8(4.7) 29.2(4.6) 29.4(4.6)LDL (mmol/L) 3.4(0.9) 3.0(0.8) 3.0(0.7) 3.0(0.7) 2.9(0.8) 2.9(0.8)HDL (mmol/L) 1.2(0.3) 1.2(0.3) 1.2(0.3) 1.2(0.3) 1.2(0.3) 1.2(0.2)Tgs (mmol/L) 1.9(1.0) 1.7(1.0) 1.6(0.8) 1.6(0.9) 1.5(0.9) 1.5(0.9)SBP (mmHg) 133.3(12.7) 129.9(13.0) 130.5(12.9) 129.7(12.1) 128.8(12.4) 125.6(11.0)DBP (mmHg) 80.5(8.6) 79.4(7.8) 79.6(7.5) 77.6(8.2) 78.6(8.4) 78.1(7.8)Insulin/body weight (IU/kg) 0.41(0.20) 0.46(0.22) 0.51(0.23) 0.52(0.24) 0.53(0.24) 0.54(0.25)% reporting hypoglycemiab 3 21.1 19.8 22.4 21.8 23.3Reported milda hypoglycaemic episodesb 15 176 168 208 214 277Reported milda episode/patient during

daytime (mean(SD))b2.2(0.8) 3.1(2.7) 3.2(1.6) 3.3(1.9) 4.0(2.9) 4.6(3.9)

Reported milda episode/patient duringnight (mean(SD))b

1(0) 1.7(1.9) 1.2(0.4) 1.9(1.2) 1.7(0.7) 1.8(1.2)

Confirmed milda hypoglycaemicepisodes/patient (mean(SD))b

1.7(0.6) 2.8(2.9) 2.3(1.3) 2.5(1.5) 3.2(3.3) 4.4(4.6)

Pt N, patients number; SD, standard deviation; FBG, fasting blood glucose; BMI, body mass index; HDL, high density lipoprotein; LDL, low densitylipoprotein; Tgs, triglycerides; SAP, systolic blood pressure; DAP, diastolic blood pressure.a “mild” are hypoglycemic episodes not needing assistance or hospitalisation.b Hypoglycaemic events at baseline, 3 months and 6 months refer to the 3-month time frame before visit. For 12, 18 and 24 months, hypogly-

caemic events refer to the 6-month time frame before visit.

p r i m a r y c a r e d i a b e t e s 7

12.83

8.267.81 7.83

7.51 7.28

9.65

7.887.44 7.39 7.28 7.14

6

7

8

9

10

11

12

13

14

Baseline 3m 6m 12m 18m 24m

Most recent FBG (mmol/l) Most recent HbA1c (%)

Number of pa�ents per visit Baseline:237, 3m:237, 6m:237, 12m:237, 18m:229, 24m:227

Fig. 2 – Glucose and HbA1c changes during the study(

o1m

ctps

3

Dfpwbmti

Fo

mean values). FBG: fasting blood glucose.

bserved at 24 months compared to baseline. HDL varied from.20(0.26) to 1.25(0.32) mmol/l. There was also a decrease inean systolic blood pressure (−7.6(14.6) mmHg) (Table 2).The proportion of patients who experienced hypogly-

aemic episodes in the 3 months before commencement ofhe study was 3%. Approximately 20% (range 19.8–23.3%) ofatients reported hypoglycaemic episodes at each consecutivetudy visit (Table 2).

.7. Health care utilisation and costs

uring the first 6 months median (IQR) total cost increasedrom 438 (230–659) to 538 (372–709)D. For the 18–24 monthseriod median costs were 451 (293–616)D (Fig. 3). Mean costsere 496(383), 573(276) and 485(247)D, respectively. Post-aseline decreases were observed for OAD cost (baseline

edian cost 197 (76–454)D, change up to 6 months −119 (−374

o −39)D and up to 24 months −130 (−426 to −59)D. Increasesn median cost were apparent for self-monitoring of blood

ig. 3 – Total direct cost – distribution of cost componentsver 6-month periods.

( 2 0 1 3 ) 235–242 239

glucose (mean (SD) number of glucose test strips used perweek was 6.5(6.6) at baseline, 8.0(6.4) at 6 months and 7.8(5.5)at 24 months at a median cost of 60 (40–140), 146 (69–305) and121 (62–223)D respectively). The mean number of total contactswith the HCP (including visits and phone calls), was 7.5 dur-ing the 6-month period immediately prior to insulin initiationand 9.8 during the 6-month period following insulin initia-tion, and 5.4, 4.8 and 4.4 between 6 and 12 months, 12 and 18months and 18 and 24 months after insulin initiation respec-tively. Primary care or specialist nurses are very rarely visitedby this Greek subsample (<5% at any given time). Regardingdiabetes complications, 36.3–57.8% of patients had visited anophthalmologist, while less than 5% had visited a podiatrist,as reported at any of the study follow-up visits. Hospitalisa-tions due to diabetes were infrequent during insulin treatment(in 7.6% of patients at baseline and 3%, 2.5%, 2.6% and 1.3% at6, 12, 18 and 24 months, respectively). Mean hospitalisationcosts are depicted in Table 3.

3.8. Compliance

The Morisky questionnaire, which explored patient compli-ance following insulin initiation, showed that percentages ofpatients reporting high compliance (score = 4) were 44.3% atinsulin initiation, 63.3% at 12 months and 71.4% at 24 months.Over the course of the study fewer patients reported that theyever forgot to take their medication (32.5% at baseline, 9.7% at24 months). Percentages of discontinuations reported becauseof feeling better (or worse) were 11.8% (28.7%) at baseline and6.6% (19.8%) at 24 months.

3.9. Quality of life

Responses regarding mobility, self-care, usual activities, pain-discomfort and anxiety-depression, accessed by the EQ-5Dquestionnaire, showed only small variations as also did theParameters of the Diabetes Health Profile [18] (barriers toactivity, excessive eating and psychological distress). Finally,results from the Experience with Insulin Therapy Question-naire [19] suggested favourable impressions from the majorityof the participants and a positive feedback regarding bloodsugar control (94.5% at 12 months and 86.3% at 24 months)and the perception of feeling better (86.9% at 12 months and80.6% at 24 months) with insulin use.

4. Discussion

Guidelines recommend that individuals with T2DM onoptimised oral glucose-lowering drugs, who experience inade-quate blood glucose control, should be offered insulin therapyearly [12,20,21] to lower HbA1c and thereby reduce the risk ofdeveloping long-term diabetic complications [1,11,12,22,23]. Inthe Hellenic subsample of the INSTIGATE study, prior to initi-ation of insulin therapy, mean HbA1c levels were well aboveInternational Diabetes Federation (IDF) guidelines [2] which

we believe is most likely due to a sustained period of poorglycaemic control. Also a high percentage of patients in thissubgroup exhibited diabetes-related co-morbidities includ-ing obesity, dyslipidaemia and hypertension. Hypertension

240 p r i m a r y c a r e d i a b e t e s 7 ( 2 0 1 3 ) 235–242

Table 3 – Median and mean costs during the study.

Periods in months 6 m prior – baseline 0–6 m 6–12 m 12–18 m 18–24 m

Median(IQR) total cost (D) 438(230–659) 538(372–709) 423(284–625) 483(291–657) 451(293–616)Mean(SD) total cost (D) 496(383) 573(276) 496(339) 504(278) 485(247)Mean(SD) cost for OADs (D) 289(294) 75(153) 65(131) 67(158) 63(133)Mean(SD) cost for insulin (D) 1(1) 252(137) 248(190) 251(155) 253(159)Mean(SD) cost for SMBG (D) 127(133) 193(147) 143(121) 158(125) 146(106)Mean(SD) cost for general practice care (D) 11(17) 13(33) 8(26) 6(12) 6(12)Mean(SD) cost for specialist care (D) 13(10) 19(15) 9(7) 10(7) 8(5)Mean(SD) total cost for HCP consultations (D) 25(16) 31(23) 18(23) 16(10) 16(12)

d blo

Mean(SD) hospitalisation cost (D) 47(187)

IQR, interquartile range; SD, standard deviation; SMBG, self-monitore

and diabetes are both known risk factors for cardiovasculardisease and together they predispose patients to coronaryartery disease, peripheral vascular disease and renal disease[24]. High HbA1c at initiation of insulin therapy (9.7%(1.6%))despite a long duration of diabetes (11.8(7.0) years) shows lateinitiation of insulin (in terms of time since diagnosis andglycaemia) which may reflect inertia, as reported in othercountries [25]. Although it is unknown on this study whetherthis inadequate initiation of insulin therapy was driven bythe healthcare providers or patients (or some combinationthereof), numerous publications have reported resistance onthe part of the patients. Reasons may include fears relatedto injectable antidiabetic treatment (such as the fear of injec-tion, hypoglycaemia or other side effects), social inhibition ifinjected publicly, negative feelings about the disease and itstreatment, and restrictions in lifestyle [26–28]. Such inertiaalso impacts patient encounters and adherence to medica-tions [29]. Unfortunately reasons for insulin initiation (andthe lateness thereof in many patients) cannot be establishedfrom INSTIGATE. In view of a high percentage of patientsnot taking any oral antidiabetic drugs at baseline (Fig. 1),while 4 weeks prior to initiation this percentage was merely2.3% [30], presumably more HCPs in Greece tended to stop allOADs when introducing patients to insulin treatment, whilethis 2.3% may be newly diagnosed patients with severelyuncontrolled diabetes, initiating insulin right at diagnosis.Also of interest is that one in ten patients reported no useof insulin at 24 months which could be interpreted as a tem-poral use of insulin to manage uncontrolled diabetes untilother therapeutic interventions (like diet or exercise) providedbenefit.

Considering insulin regimens, greater HbA1c reduction isshown when using biphasic or prandial insulin rather thana basal-only regimen [31–33]. In this study, HCPs were freeto choose the starter regimen that felt was most appropriatefor their patients. We observed a variety of starting regimens,with basal insulin prescribed most frequently (consistent withADA/EASD guidelines). Basal-bolus regimens were initiatedmore frequently than basal only or mixtures in patients witha longer duration of diabetes. Insulin treatment was initiatedwith high baseline HbA1c levels. It is of interest that basalor premixed formulations were most commonly initiated insouthern countries (Greece and Spain) while prandial insulin

was initiated in Germany, where the patients were slightlyyounger and had shorter diabetes duration when initiatinginsulin [34].

15(104) 18(158) 9(62) 6(56)

od sugar.

From the analysis of the INSTIGATE total data [14], linearregression of factors associated with change in HbA1c at 6months after initiating insulin showed that for patients, over-all, the baseline HbA1c was the most significant factor. InGreece, there was a statistically significant effect of insulinregimen and baseline HbA1c, although the factors examinedexplained very little of the variability in achieved HbA1c [14].

Insulin initiation has been associated with weight gainwhich is more often observed with mixtures and basal/bolusregimens (which also provide better glycaemic control) thanbasal-only regimens [32,33]. The Greek subsample also hada higher mean weight gain in patients initiating mixturesor basal bolus regimens than those with long acting ones.Hypoglycaemic episodes increased after insulin initiation, asexpected; however, serious events were rare. It should benoted, however, that whereas the frequency of hypoglycaemicevents after insulin initiation is, at least numerically, similar,time frames were different for different visits (3-month framefor baseline and 3- and 6-month visits and 6-month frameafterwards – Table 2, footnote).

Considering patient satisfaction, insulin, when initiated,usually provides better metabolic control, which is appre-ciated by the diabetic patients. Thus ultimately patients’self-reported diabetes-related quality of life and other patientreported outcomes (PRO) either remain the same or showsome improvement [35–38]. The Greek subsample has alsorecorded neutral or even positive outcomes on quality of lifeand PRO on insulin treatment.

Diabetes-related costs were a research focus of this study.As reported by Liatis et al. [39], during the last decade, patientswith T2DM who were followed regularly by diabetes special-ists showed improved control of cardiovascular risk factorsat a cost of a considerable increase in medication expenses.Failing to adequately control blood glucose levels has signif-icant financial impact, resulting in a 59.5% increase in themean annual patient cost [40]. The single factor having thelargest impact on costs of treatment of patients with T2DMis the development of serious diabetes-related late complica-tions [41], while regarding SMBG costs, a small percentage ofdiabetic patients incur a substantial proportion of the relativecosts [42]. In this Greek subsample, costs related to variousinsulin regimens, those associated with the increased cost ofSMBG strips required during the early phases of insulin titra-

tion and, to a much lesser degree, costs of visits to diabetesspecialists were the most important health economic bur-dens. Although the initiation of insulin required close follow

e s 7

uGftaisSieri

wsaii

shmuuuttpepyan

F

T

C

MftwP

A

TAClMTZ

dm

r

p r i m a r y c a r e d i a b e t

p by the HCP during the first 6 months, the costs of visits inreece were lower than elsewhere in the EU [16], reflecting dif-

ering healthcare environments. To the authors’ perspective,he difference of Health care environments between Greecend other EU countries could be reflected in higher prevail-ng wages for HCPs elsewhere, or in reference of patients topecialists not routinely available in Greece, such as Diabetespecialist Nurses, Diabetes Educators, Podiatrists, Nutrition-

sts and Dieticians. The breakdown of total direct costs byxpenditure category varied considerably across countries,eflecting differences in resource use patterns, prices of med-cal resources, and different health care systems [43].

An increased rate of hypoglycaemia was observed, asould be expected. However, hospitalisation decreased pre-

umably because most hypoglycaemic episodes were mildnd self-managed. Thus, the hospitalisation costs were rathernsignificant contributors of the diabetes-related costs afternsulin initiation in Greece.

Limitations of the study include the lack of randomelection of investigators. Also, the cost for diabetes-relatedospitalisations was estimated based on the daily reimburse-ent fees of the endocrinology ward or the intensive care

nit. Finally, the direct cost of diabetes care may have beennderestimated. Since no information on healthcare resourcese was collected directly from patients, it is possible thathe medical resource consumption has not been entirely cap-ured. A potential bias may have been introduced as not allatients initially enrolled were part of the analysis. How-ver, we have not found any obvious indication that excludedatients might differ from the patients included into the anal-sis. All above findings should be interpreted in the context ofnon-controlled observational study and consider the smallumber of patients analysed in this subsample.

unding

he INSTIGATE study was funded by Eli Lilly and Company.

onflict of interests

.B. has received lecture fees and advisory board honorariarom Eli Lilly and Company, as well as from other pharmaceu-ical companies developing products for treatment of patientsith type 2 diabetes. K.A., S.S. and V.D. are employees of

harmaserve-Lilly (joined venture with Eli Lilly and Company).

cknowledgements

he Hellenic INSTIGATE investigator study group:damopoulou-Kostakioti E, Alexiou Z, Benroubi M,hristodoulou I, Giannakakis I, Giannakidis S, Giannakopou-

os P, Ioannidis G, Kiagas I, Koukoulis G, Kyriakopoulos K,anes C, Panos A, Pourou E, Rigas S, Satsoglou A, Skoutas D,soumalis G, Tzounas K, Vlachogiannis N, Zacharopoulou O,

ervas A.

We wish to thank Martha Davis, Matthew Reaney and Clau-ia Nicolay for their important scientific consultancy of thisanuscript.

( 2 0 1 3 ) 235–242 241

Parts of this manuscript have been presented at the 15thInternational Congress of Endocrinology and 14th EuropeanCongress of Endocrinology, Florence May 5–9, 2012.

e f e r e n c e s

[1] I.M. Stratton, A.I. Adler, H.A. Neil, et al., Association ofglycaemia with macrovascular and microvascularcomplications of type 2 diabetes (UKPDS 35): prospectiveobservational study, British Medical Journal 321 (2000)405–412.

[2] International Diabetes Federation. IDF Diabetes Atlas, fourthed., 2009. http://www.diabetesatlas.org/ (accessed 20.08.10).

[3] P.M. Clarke, P. Glasziou, A. Patel, et al., Event rates, hospitalutilization, and costs associated with major complicationsof diabetes: a multicountry comparative analysis, PLoSMedicine 7 (2010) e1000236.

[4] E.H. Wagner, N. Sandhu, K.M. Newton, et al., Effect ofimproved glycemic control on health care costs andutilization, Journal of the American Medical Association 285(2001) 181–189.

[5] L. Ettaro, T.J. Songer, P. Zhang, M.M. Engelgau, Cost-of-illnessstudies in diabetes mellitus, Pharmacoeconomics 22 (2004)149–164.

[6] A. Brown, L.R. Reynolds, D. Bruemmer, Intensive glycemiccontrol and cardiovascular disease: an update, NatureReviews Cardiology 7 (2010) 369–375.

[7] The UK Prospective Diabetes Study Group, UK prospectivediabetes study 16: overview of 6 years’ therapy of type IIdiabetes: a progressive disease, Diabetes 44 (1995)1249–1258.

[8] M. Stolar, Glycemic control and complications in type 2diabetes mellitus, American Journal of Medicine 123 (2010)S3–S11.

[9] J.S. Skylar, R. Bergenstal, R.O. Bonow, et al., Intensiveglycaemic control and the prevention of cardiovascularevents: implications of the ACCORD, ADVANCE, and VADiabetes trials: a position statement of the AmericanDiabetes Association and a scientific statement of theAmerican College of Cardiology Foundation and theAmerican Heart Association, Diabetes Care 32 (2009)187–192.

[10] J.L. Leahy, Impaired beta-cell function with chronichyperglycemic; “overworked beta-cell” hypothesis, DiabetesReviews 4 (1996) 298–319.

[11] The UK Prospective Diabetes Study Group, Intensive bloodglucose control with sulphonylureas or insulin comparedwith conventional treatment and risk of complications inpatients with type 2 diabetes (UKPDS 33), Lancet 352 (1998)837–853.

[12] S.E. Inzucchi, R.M. Bergenstahl, J.B. Buse, et al., Managementof hyperglycemia in type 2 diabetes: a patient-centeredapproach. Position statement of the American DiabetesAssociation (ADA) and the European Association for theStudy of Diabetes (EASD), Diabetologia 55 (6) (2012)1577–1596.

[13] I. Köster, L. von Ferber, P. Ihle, et al., The cost burden ofdiabetes mellitus: the evidence from Germany-the CoDiMstudy, Diabetologia 49 (2006) 1498–1504.

[14] A. Liebl, S. Jones, M. Benroubi, et al., Clinical outcomes afterinsulin initiation in patients with type 2 diabetes: 6-monthdata from the INSTIGATE observational study in five

European countries, Current Medical Research and Opinion27 (2011) 887–895.

[15] M. Costi, H. Smith, J. Reviriego, et al., Direct health care costsin patients with type 2 diabetes mellitus six months after

e t e s

242 p r i m a r y c a r e d i a b

starting insulin treatment in Spain: the INSTIGATE study,Endocrinology and Nutrition 58 (2011) 274–282.

[16] A. Liebl, L. Breitscheidel, C. Nicolay, M. Happich, Direct costsand health-related resource utilisation in the 6 months afterinsulin initiation in German patients with type 2 diabetesmellitus in 2006: INSTIGATE study, Current Medical Researchand Opinion 24 (2008) 2349–2358.

[17] P. Muntner, C. Joyce, E. Holt, Defining the minimal detectablechange in scores on the eight-item Morisky MedicationAdherence Scale, Annals of Pharmacotherapy 45 (2011)569–575.

[18] K.A. Meadows, C. Abrams, A. Sandbaek, Adaptation of theDiabetes Health Profile (DHP-1) for use with patients withType 2 diabetes mellitus: psychometric evaluation andcross-cultural comparison, Diabetic Medicine 17 (2000)572–580.

[19] T. Kohlmann, J. Moock, Patient satisfaction in diabetesmellitus—measures and results, Deutsche MedizinischeWochenschrift 133 (2008) S120–S123.

[20] H.W. Rodbard, P.S. Jellinger, Davidson, Statement by anAmerican Association of Clinical Endocrinologists/AmericanCollege of Endocrinology Consensus Pane l on type 2diabetes mellitus: an algorithm for glycemic control,Endocrine Practice 15 (2009) 540–559.

[21] National Institute for Health and Clinical Excellence ClinicalGuideline G Management of type 2 diabetes London: NICE,Ref NO1288 (2002, Revised 2005).

[22] European Diabetes Policy Group, A desktop guide to type 2diabetes mellitus, Diabetic Medicine 16 (1999) 716–730.

[23] W.C. Hsu, Consequences of delaying progression to optimaltherapy in patients with type 2 diabetes not achievingglycemic goals, Southern Medical Journal 102 (2009) 67–76.

[24] J.R. Sowers, M. Epstein, Diabetes mellitus and associatedhypertension, vascular disease, and nephropathy. Anupdate, Hypertension 26 (1995) 869–879.

[25] E.S. Moghissi, Insulin strategies for managing inpatient andoutpatient hyperglycemia and diabetes, Mount Sinai Journalof Medicine 75 (2008) 558–566.

[26] M. Brod, J.H. Kongsø, S. Lessard, T.L. Christensen,Psychological insulin resistance: patient beliefs andimplications for diabetes management, Quality of LifeResearch 18 (2009) 23–32.

[27] L.S. Phillips, W.T. Branch, C.B. Cook, et al., Clinical inertia,Annals of Internal Medicine 135 (9) (2001) 825–834.

[28] M.L. Parchman, J.A. Pugh, R.L. Romero, K.W. Bowers,Competing demands or clinical inertia: the case of elevatedglycosylated hemoglobin, Annals of Family Medicine 5 (3)(2007) 196–201.

[29] R. Grant, A.S. Adams, C.M. Trinacty, et al., Relationshipbetween patient medication adherence and subsequent

clinical inertia in type 2 diabetes glycemic management,Diabetes Care 30 (4) (2007) 807–812.

[30] S. Jones, M. Benroubi, C. Castell, et al., Characteristics ofpatients with type 2 diabetes mellitus initiating insulin

7 ( 2 0 1 3 ) 235–242

therapy: baseline data from the INSTIGATE study, CurrentMedical Research and Opinion 25 (3) (2009) 691–700.

[31] J. Rosenstock, Basal insulin supplementation in type 2diabetes; refining the tactics, American Journal of Medicine116 (2004) 10S–16S.

[32] D.S. Lasserson, P. Glasziou, R. Perera, et al., Optimal insulinregimens in type 2 diabetes mellitus: systematic review andmeta-analyses, Diabetologia 52 (2009) 1990–2000.

[33] M.J. Van Avendonk, G.E. Rutten, Insulin therapy in type 2diabetes: what is the evidence? Diabetes, Obesity andMetabolism 11 (2009) 415–432.

[34] A. Liebl, S. Jones, M. Benroubi, et al., Clinical outcomes afterinsulin initiation in patients with type 2 diabetes: 6-monthdata from the INSTIGATE observational study in fiveEuropean countries, Current Medical Research and Opinion27 (5) (2011) 887–895.

[35] R. Qayyum, L. Greene, AHRQ’s comparative effectivenessresearch on premixed insulin analogues for adults with type2 diabetes: understanding and applying the systematicreview findings, Journal of Managed Care Pharmacy 17 (2011)S3–S19.

[36] H. Ishii, M. Iwase, H. Seino, et al., Assessment of quality oflife in patients with type 2 diabetes mellitus before and afterstarting biphasic insulin aspart 30 (BIAsp 30) therapy:IMPROVE study in Japan, Current Medical Research andOpinion 27 (2011) 643–650.

[37] A. Braun, A. Sämann, T. Kubiak, et al., Effects of metaboliccontrol, patient education and initiation of insulin therapyon the quality of life of patients with type 2 diabetesmellitus, Patient Education and Counseling 73 (2008) 50–59.

[38] L.J. Lee, J.L. Fahrbach, L.M. Nelson, Effects of insulininitiation on patient-reported outcomes in patients withtype 2 diabetes: results from the durable trial, DiabetesResearch and Clinical Practice 89 (2012) 157–166.

[39] S. Liatis, P. Thomakos, S. Papaoikonomou, et al., Trends inthe management of type 2 diabetes and its prescription drugcosts in Greece (1998 & 2006), Experimental and ClinicalEndocrinology and Diabetes 117 (2009) 505–510.

[40] K. Athanasakis, M. Ollandezos, A. Angeli, et al., Estimatingthe direct cost of type 2 diabetes in Greece: the effects ofblood glucose regulation on patient cost, Journal of DiabetesMedicine 27 (2010) 679–684.

[41] B. Jönsson, Revealing the cost of Type II diabetes in Europe,Diabetologia 45 (2002) S5–S12.

[42] R.L. Kjome, A.G. Granas, K. Nerhus, et al., The prevalence ofself-monitoring of blood glucose and costs of glucometerstrips in a nationwide cohort, Diabetes Technology andTherapeutics 12 (2010) 701–705.

[43] S. Jones, C. Castell, A. Goday, et al., Increase in directdiabetes-related costs and resource use in the 6 months

following initiation of insulin in patients with type 2diabetes in five European countries: data from theINSTIGATE study, Clinicoeconomics and Outcomes Research4 (2012) 383–393.