Rutten_Q8_Layout 1 27/08/2009 16:31 Page 33
Screening for Type 2 Diabetes – The ADDITION Netherlands Study
The study was approved by the medical ethical committee of the to determine the extent of healthcare utilisation and to evaluate the
University Medical Centre Utrecht. Participants gave written occurrence of CVD. Patients with diabetes from practices in the
informed consent. intervention arm of the ADDITION study were excluded. We
collected data at baseline and after three years from 354 subjects in
General Practitioners and Practices four categories of glucose regulation: type 2 diabetes (n=64),
A total of 106 GPs in 79 practices participated in the study; 41 impaired glucose tolerance (IGT) (n=62), impaired fasting glucose
practices participated in the first screening procedure and 38 in the (IFG) (n=86) and normal glucose tolerance (NGT) but with an
second. Forty-two practices were randomised to the routine care elevated risk score (n=142).
group and 37 to the intensified treatment group.
Measurements
Screening Programme Capillary blood glucose values were determined using a HemoCue
Two stepwise screening procedures were performed. The first B-Glucose Analyzer based on the glucosedehydrogenase method.
procedure (May 2002 – January 2003) consisted of four steps: a self- Plasma glucose was measured using a peroxidase method. HbA
1c
completed risk questionnaire, random glucose measurement (RBG), was assessed with high-performance liquid chromatography (160 A.
fasting glucose measurement (FBG) and an oral glucose tolerance test Menarini Diagnostics). Cholesterol, high-density lipoprotein (HDL)
(OGTT). In the second procedure (July 2003 – April 2004), no RBG was cholesterol and triglycerides were determined using enzymatic
performed. The initial questionnaire contained questions about age, techniques (LX20 Beckman). The Short Form (SF)-36 measures general
gender, body mass index (BMI), family history of diabetes, frequent health, vitality, mental health, physical functioning, limitations due to
thirst, the use of antihypertensive medication, shortness of breath, physical difficulties (role physical), bodily pain, social functioning and
claudication and cycling.
7
People who scored above a pre-defined limitations due to emotional difficulties (role emotional). For each
threshold were invited to undergo subsequent diagnostic glucose dimension, scores range from 0 (worst health) to 100 (best health).
testing. Participants were classified according to the 1999 World
Health Organization (WHO) criteria.
8
Data Analyses
The association between risk score and different glucose tolerance
Intensified Multifactorial Treatment categories was analysed using linear regression. Associations of the
The intensified treatment was according to the common ADDITION yield of screening with GP and practice characteristics were studied
treatment protocol, which is characterised by intensive treatment of using multiple regression analysis. An independent variable in the
glucose, blood pressure and lipids and structured lifestyle education model was the number of detected patients with diabetes per
(dietary modification, weight loss, increased physical activity, smoking practice after adjustment for practice size and age distribution
cessation and improving adherence to medication). The glycated (standardised practice). Changes in biomedical variables after one
haemoglobin (HbA
1c
) level should be kept under 7.0%. Intensification year were compared between both groups using generalised
of glucose-lowering therapy should be initiated while HbA
1c
is >6.5%. estimating equations adjusting for age, gender, baseline values and
If HbA
1c
remains above 7.0% with oral agents, insulin therapy should clustering at practice level. Analyses were based on intention-to-
be initiated. Antihypertensive agents were prescribed if blood treat. SF-36 scores at the end of study were analysed using analysis
pressure was >120/80mmHg (systolic or diastolic). Angiotensin- of co-variance (ANCOVA), with age, gender, and baseline values as
converting enzyme (ACE) inhibitors were the initial treatment. In the co-variates to adjust for differences between treatment groups at
case of side effects, an angiotensin-II receptor antagonist was randomisation. Paired-sample t-tests were used to compare baseline
prescribed. If blood pressure was >135/85 mmHg, the dose had to be with follow-up within each treatment group. Gender and age at
increased, and thiazides, calcium channel blockers or beta-blockers baseline were compared between the different diagnostic categories
were added using a stepwise approach. Treatment with a statin was using chi-square test for dichotomous variables and analysis of
indicated if cholesterol was >5.0mmol/l or >4.5mmol/l in patients variance (ANOVA) for continuous variables, respectively. Healthcare
without or with a known history of CVD, respectively. The dose of encounters were analysed using the Mann-Whitney test. Differences
statin was increased up to maximum if cholesterol remained above between diagnostic categories in the use of medication were
threshold. In 2003, the protocol changed: all participants with examined with the chi-square test. Cardiovascular events and death
cholesterol >3.5mmol/l were treated with a statin. Acetylsalicylic acid were calculated per diagnostic category per 1,000 person-years of
80mg was given to patients treated with antihypertensive agents. In follow-up. Cox regression analysis with age and gender entered as
the routine care group, the treatment goals according to the Dutch co-variates in the model was used to examine whether the diagnostic
guidelines were less ambitious. The target value for HbA
1c
was <7.0%; categories have different risks regarding mortality and cardiovascular
however, HbA
1c
levels between 7.0 and 8.5% were described as morbidity. The data were analysed applying the SPSS statistical
acceptable. Lipid-lowering drugs were initiated if participants without package (version 15.0). A p-value <0.05 was considered significant.
a known history of CVD had a greater than 25% risk of developing a
coronary disease within 10 years, while all those with previous CVD Results
were treated with a statin. If blood pressure was >150/85mmHg, The screening algorithm and yields are presented in Figures 1 and 2.
treatment was recommended. Detailed instructions about providing Eventually, 586 participants (1.0%) were diagnosed with type 2
lifestyle education were not given. diabetes (four-step procedure: 285 people; three-step procedure: 301
people). Impaired glucose regulation was diagnosed in 1,011
Three-year Follow-up of Persons in participants (1.8%).
9
Different Glucose Regulation Categories
All persons with a risk score above threshold from 24 practices, Risk scores (± standard deviation [SD]) of persons with NGT, IFG, IGT,
screened from May to October 2002, were followed over three years epidemiological diabetes and type 2 diabetes in the four-step
EUROPEAN ENDOCRINOLOGY 33