Received: 23 March 2017 Revised: 29 September 2017 Accepted: 5 November 2017

DOI: 10.1002/dmrr.2968

RESEARCH ARTICLE

Early‐onset type 2 diabetes: Age gradient in clinical and

behavioural risk factors in 5115 persons with newly diagnosed
type 2 diabetes—Results from the DD2 study
A. Bo1,2 | R.W. Thomsen3 | J.S. Nielsen4 | S.K. Nicolaisen3 | H. Beck‐Nielsen4 |

J. Rungby5,6 | H.T. Sørensen3 | T.K. Hansen7 | J. Søndergaard8 | S. Friborg9 |

2 2,10
T. Lauritzen | H.T. Maindal

1
Danish Diabetes Academy, Odense, Denmark
2
Abstract
Department of Public Health, Aarhus
University, Aarhus, Denmark Aim: To examine the association between early onset of type 2 diabetes mellitus (DM) and
3
Department of Clinical Epidemiology, Aarhus clinical and behavioural risk factors for later complications of diabetes.
University Hospital, Aarhus, Denmark
4
Methods: We conducted a cross‐sectional study of 5115 persons with incident type 2 DM
Diabetes Research Centre, Department of
Endocrinology, Odense University Hospital, enrolled during 2010‐2015 in the Danish Centre for Strategic Research in Type 2 Diabetes‐
Odense, Denmark cohort. We compared risk factors at time of diagnosis among those diagnosed at ≤45 years (early
5
Department of Biomedicine, Aarhus onset) with diagnosis age 46 to 55, 56 to 65 (average onset = reference), 66 to 75, and >75 years
University Hospital, Aarhus, Denmark (late onset). Prevalence ratios (PRs) were computed by using Poisson regression.
6
Center for Diabetes Research, Gentofte
University Hospital, Copenhagen, Denmark Results: Poor glucose control, ie, HbA1c ≥ 75 mmol/mol (≥9.0%) in the early‐, average‐, and
7
Department of Internal Medicine and
late‐onset groups was observed in 12%, 7%, and 1%, respectively (PR 1.70 [95% confidence
Endocrinology, Aarhus University Hospital, intervals (CI) 1.27, 2.28] and PR 0.17 [95% CI 0.06, 0.45]). A similar age gradient was observed
Aarhus, Denmark for severe obesity (body mass index > 40 kg/m2: 19% vs. 8% vs. 2%; PR 2.41 [95% CI 1.83,
8
General Practice Research Unit, Department 3.18] and 0.21 (95% CI 0.08, 0.57]), dyslipidemia (90% vs. 79% vs. 68%; PR 1.14 [95% CI 1.10,
of Public Health, University of Southern
1.19] and 0.86 [95% CI 0.79, 0.93]), and low‐grade inflammation (C‐reactive protein > 3.0 mg/L:
Denmark, Odense, Denmark
9 53% vs. 38% vs. 26%; PR 1.41 [95% CI 1.12, 1.78] and 0.68 [95% CI 0.42, 1.11]). Daily smoking
Department of Endocrinology, Odense
University Hospital, Odense, Denmark was more frequent and meeting physical activity recommendations less likely in persons with
10
Steno Diabetes Centre Copenhagen, Health early‐onset type 2 DM.
Promotion, Gentofte, Denmark
Conclusions: We found a clear age gradient, with increasing prevalence of clinical and behav-
Correspondence
ioural risk factors the younger the onset age of type 2 DM. Younger persons with early‐onset
Anne Bo, Department of Public Health,
Bartholins alle 2, 8000 Aarhus C, Aarhus, type 2 DM need clinical awareness and support.
Denmark.
Email: anne.bo@ph.au.dk KEY W ORDS
Funding information cross sections study, DD2 study, early onset type 2 diabetes, health behaviour, risk factors
Novo Nordisk A/S; Aarhus University; Riisfort
Foundation; Danish Health and Medicines
Authority; Danish Diabetes Association;
Central Denmark Region; Danish Agency for
Science, Grant/Award Numbers: 09‐067009
and 09‐075724; Danish Diabetes Academy;
Novo Nordisk Foundation

An abstract with similar but less detailed analysis has been presented with a 1 | I NT R O DU CT I O N
poster at the European Association for the Study of Diabetes (EASD) 2016:
Bo, A et al. High burden of cardiovascular risk factors and poor glycemic control
The rising global burden of type 2 diabetes mellitus (DM) in middle‐
in type 2 diabetes patients diagnosed before the age of 45 years in Denmark:
results from the Danish Center for Strategic research in Type 2 Diabetes (DD2) aged and older persons is now accompanied by increasing prevalence
study. 52nd EASD Annual Meeting, Munich, 12‐16 September 2016. in youth and younger adults.1-5 This is alarming because early‐onset

Diabetes Metab Res Rev. 2018;34:e2968. wileyonlinelibrary.com/journal/dmrr Copyright © 2017 John Wiley & Sons, Ltd. 1 of 9
https://doi.org/10.1002/dmrr.2968
2 of 9 BO ET AL.

type 2 DM (onset age < 40 or 45 years) is likely to be associated with himself or may refer the patient to a hospital outpatient clinic, where
increased risk of complications later in life. Epidemiological studies all procedures are performed. Upon enrolment, an online registration
have shown increased rates of retinopathy, nephropathy, cardiovascu- form21 containing patient‐reported and clinical examination data is
lar disease, and premature mortality among persons with early‐onset completed. Fasting urine and blood samples are obtained and then
6-11
type 2 DM, compared with persons with later‐onset type 2 DM. stored in the DD2 biobank.
The excess burden of late complications is likely related to longer dis- The unique Central Personal Registration number provided to all
ease duration during the life course among early‐onset cases but may Danish residents at birth or upon immigration is used to link DD2 data
also be due to higher prevalence of risk factors at the time of type 2 with Danish national health registries. In our study, these included (1)
12
DM onset. Cross‐sectional studies have reported a higher prevalence the Danish National Prescription Registry, containing individual‐level
of poor glucose control, obesity, hypertension, increased low‐density information on prescriptions dispensed from all Danish community
lipoprotein (LDL), and of family history of type 2 diabetes among per- pharmacies; (2) the Danish National Patient Registry, containing infor-
sons with early‐onset type 2 DM compared with later‐onset individ- mation on hospital inpatient and outpatient clinic contacts; and (3) the
13-19
uals at different duration of type 2 DM. The few studies of Danish Diabetes Database for Adults (DDDA), a nationwide quality‐of‐
behavioural risk factors in persons with early‐ vs. later‐onset type 2 care database containing indicators for adults with diabetes reported
DM have reported a higher smoking prevalence and lower physical from general practices and hospital outpatient clinics.22 Supplementary
16,20
activity level among early‐onset individuals. Former studies pri- data sources in the DD2 have been described by Thomsen et al.22
marily used a dichotomization of “early” (at <40 or 45 years of age) ver- Fasting glucose was measured as a part of the enrolment proce-
sus “late” onset (at >40 or 45 years of age). This may hide possible age dure. Information on HbA1c was collected from the DDDA, using the
differences in presence of clinical and behavioural risk factors, which HbA1c value measured closest to the DD2 enrolment date. The
can be crucial for providing the appropriate health services to persons chosen cut point for increased LDL cholesterol at 2.5 mmol/L is
newly diagnosed with type 2 DM. the Danish recommended threshold for initiating lipid‐lowering
In the present study, we obtained information at time of diagnosis treatment.23 We defined the presence of any dyslipidemia
for a large cohort of persons clinically diagnosed with type 2 DM. We according to the American Diabetes Association24: LDL cholesterol
hypothesized that persons with early‐onset type 2 DM have a high >2.60 mmol/L, or HDL cholesterol <1.02 mmol/L, or triglyceride
burden of clinical and behavioural risk factors for later complications. >1.7 mmol/L. C‐reactive protein (CRP) was available for a biobank
We aimed to explore whether there is a gradient in the association subgroup of the first consecutive 1037 patients enrolled in the DD2
between age at type 2 DM diagnosis and prevalence of these risk project. Glutamic Acid Decarboxylase (GAD) antibody was measured
factors. after enrolment, based on biobank samples. Data on behavioural risk
factors at enrolment included physical activity, alcohol consumption,
and smoking. Information on physical activity and alcohol consumption
2 | METHODS was self‐reported. The assessment of physical activity level was based
on number of days with at least 30 minutes of moderate to hard
2.1 | Study design and study population physical activity. High‐risk alcohol consumption was categorized

In a cross‐sectional study based on a cohort of persons newly diagnosed according to the Danish Health Authority's definitions as more

with type 2 DM, we compared clinical and behavioural risk factors among than 21 and 14 drinks weekly for men and women, respectively.

persons diagnosed at age ≤ 45 years (early onset) and those diagnosed at Information on smoking was obtained from the DDDA, and informa-

ages 46 to 55 years, 56 to 65 years (average onset = reference group), 66 tion on medications at enrolment was obtained from the Danish

to 75 years, and >75 years (late onset). A total of 7053 participants were National Prescription Registry.

enrolled consecutively from general practices and hospital outpatient

clinics between January 2010 and June 2015 as part of a nationwide
2.3 | Ethics
cohort established by the Danish Center for Strategic research in Type
2 Diabetes (DD2).21 Of these participants, we were able to include Participants in the DD2 project signed a written informed consent doc-

5115 patients who could currently be linked to other databases for a ument, after receiving information approved by the Danish National

detailed assessment of risk factors (see below). Committee on Health Research Ethics. Patient registration and
biospecimen collection for the DD2 project were approved by the
Danish National Committee on Health Research Ethics (record number
2.2 | Data collection S‐20100082) and the Danish Data Protection Agency (record number
Enrolment in the DD2 cohort has previously been described by Nielsen 2008‐58‐0035).
21
et al. In brief, patients are diagnosed with type 2 DM in everyday
routine clinical practice—either by hospital physicians or general
practitioners—and are thereafter invited to participate in the DD2 pro-
2.4 | Statistical analysis
ject. In both settings, the diagnostic criteria have followed Danish Prevalence proportions were estimated for categorical variables.
national guidelines and World Health Organization criteria throughout Medians with interquartile ranges were calculated for continuous
the study period. If a patient gives informed consent to participate, the variables, as data were not normally distributed. To compare
physician may choose to perform the DD2 enrolment procedures prevalence in different age groups, prevalence ratios (PRs) with 95%
BO ET AL. 3 of 9

confidence intervals (95% CIs) were calculated by using Poisson 3.3 | Other clinical risk factors
regression analysis. The average‐onset group (diagnosis at age 56‐
In the early‐onset group, 39% had BMI >35, 9 in 10 had central obesity
65 years) was used as the reference group. For many patients, the first
(88%), one third had hypertension (32%), and close to half had a LDL
prescription of a glucose lowering drug (GLD) had been made several
>2.5 mmol/L (50%) and CRP >3.0 mg/L (53%) (Table 2). For risk factors
months before enrolment into the DD2 cohort. Moreover, the treat-
such as high BMI, LDL, and CRP, there was a clear gradient of higher
ment duration before enrolment varied across age groups (Table 4),
prevalence with earlier age of onset (Figures 1C and 1E). For example,
and therefore could be considered a possible confounder for parame-
the gradient for BMI >40 was 19% (early onset), 8% (average onset),
ters affected by GLDs. For this reason, we adjusted the PRs of body
and 2% (late onset) (PR 2.41 [95% CI 1.83, 3.18] and 0.21 [95% CI
mass index (BMI), central obesity, HbA1c, and fasting plasma glucose
0.08, 0.57]). The gradient for dyslipidemia was 90% vs. 79% vs. 68%
for time elapsed between first dispensed prescription of a GLD and
for the three groups (PR 1.14 [95% CI 1.10, 1.19] and 0.86 [95% CI
enrolment in the DD2 cohort. We refrained from using further multi-
0.79, 0.93]), and for CRP >3.0 mg/L, this was 53%, 38%, and 26%
variate adjustment models, because most of the risk factors we exam-
(PR 1.41 [95% CI 1.12, 1.78] and 0.68 [95% CI 0.42, 1.11]) (Table 2).
ined may act as intermediates and clusters in the same causal
The prevalence of hospital‐diagnosed retinopathy was also higher in
pathophysiological pathways and are impossible to disentangle in a
the early‐onset group (7%) than in the average‐onset group (5%) (PR
cross‐sectional design. All analyses were performed by using SAS ver-
1.58 [95% CI 1.08, 2.31]). Albuminuria was present in close to one in
sion 9.2 (SAS institute, Inc., Cary, North Carolina).
five persons in the four youngest onset groups and still higher in the
late‐onset group (28%) (PR for onset >75 years vs. 56‐65 years: 1.44
[95% CI 1.18, 1.76]). Hypertension was present in about one third of
3 | RESULTS
persons in all five age‐of‐onset groups (Table 2 and Figure 1D).

3.1 | Background characteristics

Median age in the study cohort was 62.2 years (interquartile range:
3.4 | Behavioural risk factors
53.3‐68.8 years), and 10% were diagnosed at age ≤ 45 years (early The prevalence of performing at least 30 minutes of moderate to
onset). The proportion of women was approximately 40% in the <45, intensive physical activity/day was 22% in the early‐onset group,
46 to 55, and 66 to 75 year age groups and 49% in the late‐onset increasing with age to 32% in the average‐onset group (PR for early
(>75 years) group (Table 1). Family history of type 2 DM was clearly vs. average onset: 0.69 [95% CI 0.58, 0.82]), peaking at 36% among
highest in the early‐onset group (64%), decreasing to 52% in the aver- 66 to 75 year olds (PR for the 66‐75 years‐group vs. average‐onset
age‐onset group, and to 40% in the late‐onset group. The prevalence group: 1.12 [95% CI 1.02, 1.23]) (Table 3 and Figure 1F). There was
of previous gestational diabetes mellitus was 22% among women in also a gradient in prevalence of daily smoking; 24%, 20%, and 8% in
the early‐onset group—much higher than in the other groups. A posi- the early‐, average‐, and late‐onset groups, respectively (PR 1.18
tive test for GAD antibodies was present in 7% in the early‐onset [95% CI 0.98, 1.42] and PR 0.39 [95% CI 0.27, 0.56]) (Table 3 and
group, and in less than 3% in the age groups above 46 years (PR for Figure 1F). Self‐reported high‐risk alcohol intake was below 10% in
early onset vs. average onset 2.60 [95% CI 1.66; 4.08]) (Table 1). While all age groups and lower in the early‐onset (4%) than in the average‐
positive GAD antibodies in retrospect indicates autoimmune diabetes, onset (7%) group (PR 0.50 [95% CI 0.31, 0.81]) (Table 3).
we kept these few patients in our cohort for reasons of completeness
and generalizability for everyday clinical practice‐diagnosed type 2 DM
(when re‐running all analysis while excluding GAD positive patients, all 3.5 | Therapy
results only changed marginally [data not shown]).
Between 75% and 80% of persons in the five age‐of‐onset groups
received noninsulin GLDs at time of cohort enrolment (Table 4). The
prevalence of using no GLDs was lowest in the early‐onset group
3.2 | Glucose control (8%) and increased to 15% and 22% in the average‐ and late‐onset
We observed a clear gradient of increasingly poor glucose control in groups, respectively (PR for early onset 0.56 [95% CI 0.41, 0.76] and
members of the younger‐onset groups, after adjusting for duration of for late onset 1.47 [95% CI 1.18, 1.84]). In contrast, the use of both
glucose‐lowering treatment before enrolment (Table 1 and Figures 1A insulin and noninsulin drugs was highest in the early‐onset group
and 1B). For example, the prevalence of poor control, ie, HbA1c (11%) and decreased to 6% and 2% in the average‐ and late‐onset
≥75 mmol/mol (≥9.0%) in the early‐, average‐, and late‐onset groups, groups. (PR 1.78 [95% CI 1.30, 2.45] and PR 0.39 [95% CI 0.20, 0.77])
was observed in 12%, 7%, and 1%, respectively (PR 1.70 [95% CI (Table 4 and Figure 1G). Insulin use was higher (21%) among GAD anti-
1.27, 2.28] and PR 0.17 [95% CI 0.06, 0.45]) (Table 1). A reverse gradi- body positive cohort members than among the GAD antibody negative
ent that was found for good glucose control, ie, HbA1c 48 to 53 mmol/ members (7%), but the finding of higher insulin use in the early‐onset
mol (6.5%‐7.0%) in the early‐, average‐, and late‐onset groups, was group remained when excluding GAD antibody positive individuals
observed in 20%, 27%, and 35%, respectively (PR 0.73 [95% CI 0.60, (data not shown). In the early‐onset group, 40% received antihyperten-
0.89] and PR 1.28 [95% CI 1.08, 1.51]) (Table 1). A similar age gradient, sive drugs, 53% received lipid‐lowering drugs, and 8% received
with higher glucose levels in younger age groups, was observed for anticoagulation drugs. Use of all three drug types increased with age
fasting plasma glucose (Figure 1B). of onset (Figure 1H).
4 of 9 BO ET AL.

TABLE 1 Background characteristics and glucose control among 5115 persons with newly diagnosed type 2 diabetes in 5 groups defined by age at
diagnosis
Age at Diagnosis
Missing ≤45 years 46‐55 years 56‐65 years 66‐75 years >75 years
Background Characteristics N n = 516 n = 1091 n = 1651 n = 1466 n = 391

Female, % 0 42.2 39.5 42.0 40.9 48.6

PR (95% CI) 1.01 (0.90‐1.13) 0.94 (0.86‐1.03) 1 (ref) 0.97 (0.90‐1.06) 1.16 (1.03‐1.30)
Family history of T2DM I,% 0 63.6 61.8 51.9 46.0 40.2
PR (95% CI) 1.23 (1.13‐1.33) 1.19 (1.11‐1.27) 1 (ref) 0.89 (0.83‐0.95) 0.77 (0.68‐0.88)
Previous gestational diabetesII, % 0 21.6 3.7 0.1 ‐ ‐
III
GAD‐positive , % 270 6.8 3.0 2.6 2.3 0.3
PR (95% CI) 2.60 (1.66‐4.08) 1.14 (0.72‐1.81) 1 (ref) 0.87 (0.55‐1.38) 0.10 (0.01‐0.74)
Enrolment performed in hospital outpatient 0 64.1 56.7 51.2 40.9 28.9
clinic, %
Enrolment performed in general practice, % 0 35.9 43.3 48.8 59.1 71.1
Glucose Control
HbA1c, mmol/mol, m (IQR)* 83 6.8 (6.2‐8.0) 6.7 (6.2‐7.4) 6.6 (6.1‐7.2) 6.4 (6.1‐6.9) 6.5 (6.1‐6.9)
<48 (<6.5%), % 35.5 40.0 43.9 50.1 48.4
PR (95% CI) 0.85 (0.74‐0.97) 0.93 (0.85‐1.03) 1 (ref) 1.14 (1.05‐1.24) 1.15 (1.01‐1.31)
48‐53 (6.5‐7%), % 20.4 24.8 26.7 29.3 34.8
PR (95% CI) 0.73 (0.60‐0.89) 0.90 (0.79‐1.03) 1 (ref) 1.11 (0.99‐1.24) 1.28 (1.08‐1.51)
53‐58 (7.0‐7.5%), % 12.0 11.9 11.7 9.1 11.8
PR (95% CI) 0.99 (0.76‐1.3) 1.02 (0.82‐1.25) 1 (ref) 0.80 (0.65‐0.99) 0.97 (0.70‐1.33)
58‐75 (7.5‐9.0%), % 19.8 13.9 10.8 9.1 3.9
PR (95% CI) 1.75 (1.40‐2.19) 1.26 (1.02‐1.54) 1 (ref) 0.86 (0.70‐1.07) 0.38 (0.23‐0.64)
≥75 (≥9.0%), % 12.4 9.3 6.8 2.4 1.0
PR (95% CI) 1.70 (1.27‐2.28) 1.33 (1.03‐1.73) 1 (ref) 0.37 (0.25‐0.54) 0.17 (0.06‐0.45)
Fasting blood glucose, mmol/L* 637 7.6 (6.5‐9.2) 7.4 (6.5‐8.7) 7.2 (6.4‐8.3) 7.0 (6.3‐7.9) 6.8 (6.2‐7.7)
<6.5, % 25.1 25.7 28.3 31.8 38.3
PR (95% CI) 0.88 (0.73‐1.06) 0.92 (0.79‐1.06) 1 (ref) 1.12 (0.98‐1.26) 1.43 (1.21‐1.70)
6.5‐7.0, % 11.5 14.5 16.8 19.1 18.7
PR (95% CI) 0.73 (0.54‐0.97) 0.90 (0.73‐1.10) 1 (ref) 1.20 (1.01‐1.43) 1.23 (0.95‐1.60)
7.0‐7.5, % 11.8 13.2 14.3 14.3 15.4
PR (95% CI) 0.82 (0.61‐1.10) 0.93 (0.74‐1.15) 1 (ref) 1.02 (0.84‐1.23) 1.02 (0.75‐1.38)
7.5‐9.0, % 24.2 25.7 24.6 23.0 21.2
PR (95% CI) 0.96 (0.80‐1.17) 1.01 (0.87‐1.16) 1 (ref) 0.93 (0.81‐1.07) 0.81 (0.64‐1.03)
≥9.0, % 27.4 20.9 16.1 11.8 6.4
PR (95% CI) 1.65 (1.36‐2.00) 1.28 (1.08‐1.52) 1 (ref) 0.73 (0.60‐0.88) 0.41 (0.27‐0.62)

Estimates shown as percentages (%), medians (m) with interquartile ranges (IQR), or prevalence ratios (PR) with 95% confidence Intervals (95% CI).
*Adjusted for time since commencement of glucose‐lowering treatment to enrolment in the DD2 cohort.
I
Cohort member reported a father, mother, or child with type 2 diabetes.
II
Estimates are given only as percentage of women.
III
Glutamic Acid Decarboxylase antibody levels >30 kU/l.

4 | DISCUSSION microalbuminuria were present at worrisome levels among persons

with early‐onset type 2 DM in light of their short disease duration
The key finding in this study was an age gradient of increasingly and young age.
higher prevalence of most clinical and behavioural risk factors with Previous studies corroborate clustering of risk factors in persons
younger age at type 2 DM diagnosis. Persons with early‐onset type with newly diagnosed early‐onset type 2 DM. A US study18 found a
2 DM had a markedly higher prevalence of severe obesity, dyslipid- higher mean BMI [39 kg/m2 vs. 33 kg/m2, P < .001] and higher mean
emia, low‐grade inflammation, tobacco smoking, and physical inactiv- HbA1c level (61 mmol/mol [7.7%] vs. 58 mmol/mol [7.5%], P = .030)
ity. Importantly, persons in the early‐onset group also had poorer in persons aged <45 years at diagnosis compared with those >45 years,
glucose control than persons in the later‐onset groups, although and the study found equally high prevalence of abnormal lipids [82%
the early‐onset individuals were more likely to receive both insulin vs. 78%, P = .130] in the 2 groups. A UK study25 reported a mean
and noninsulin glucose lowering treatment. Retinopathy and BMI of 33 kg/m2 and identified 80% with HbA1c >53 mmol/mol
BO ET AL. 5 of 9

FIGURE 1 Prevalence of risk factors for diabetes related complications among 5115 people with newly diagnosed type 2 diabetes in 5 groups
defined by age at diagnosis

(7%) and 37% with hypertension among persons aged <40 years with associated with greater increase in HbA1c after type 2 DM
newly diagnosed type 2 DM. Studies of persons with longer disease diagnosis,26 and other studies found that persons with early‐onset
durations also found a worse risk factor profile among persons with type 2 DM more often progress to using insulin, but remain more
early‐onset type 2 DM than among later‐onset individuals.2,16 A poorly controlled, than persons with later‐onset type 2 DM.7,16,17 A
large10‐year follow‐up study showed that that younger age is recent DD2 study found that young age was associated with a 1.3‐fold
6 of 9 BO ET AL.

TABLE 2 Clinical risk factors among 5115 persons with newly diagnosed type 2 diabetes in 5 groups defined by age at diagnosis
Age at Diagnosis
Missing ≤45 years 46‐55 years 56‐65 years 66‐75 years >75 years
Clinical Risk Factors N n = 516 n = 1091 n = 1651 n = 1466 n = 391

Body mass index, kg/m2, m (IQR)* 1280 32.7 (28.7‐37.8) 31.7 (27.9‐35.9) 30.8 (27.1‐34.5) 29.4 (26.3‐32.9) 28.4 (25.5‐31.2)

BMI <25, % 6.9 9.9 13.4 15.8 20.3

PR (95% CI) 0.53 (0.36‐0.77) 0.71 (0.55‐0.92) 1 (ref) 1.19 (0.96‐1.46) 1.48 (1.09‐2.00)

BMI 25‐30, % 26.6 30.0 31.5 39.2 45.1

PR (95% CI) 0.86 (0.72‐1.03) 0.96 (0.84‐1.10) 1 (ref) 1.23 (1.10‐1.39) 1.42 (1.20‐1.68)
BMI 30‐35, % 27.8 31.4 31.5 28.5 26.7

PR (95% CI) 0.87 (0.72‐1.03) 1.01 (0.88‐1.15) 1 (ref) 0.91 (0.80‐1.03) 0.88 (0.70‐1.10)
BMI 35‐40, % 19.4 17.6 15.7 12.7 6.4

PR (95% CI) 1.22 (0.96‐1.56) 1.12 (0.92‐1.37) 1 (ref) 0.83 (0.67‐1.02) 0.44 (0.27‐0.71)
BMI >40, % 19.4 11.1 8.0 3.9 1.5

PR (95% CI) 2.41 (1.83‐3.18) 1.37 (1.04‐1.80) 1 (ref) 0.51 (0.36‐0.74) 0.21 (0.08‐0.57)
Waist‐to‐hip ratio, m (IQR) 4 0.97 (0.92‐1.03) 1.00 (0.98‐1.03) 1 (ref) 0.98 (0.96‐1.00) 0.92 (0.88‐0.96)
Central obesityI, % 87.8 92.8 92.5 90.6 84.9
PR (95% CI) 0.95 (0.92‐0.98) 1.00 (0.98‐1.03) 1 (ref) 0.98 (0.96‐1.00) 0.92 (0.87‐0.96)
Weight gain since age 20 years, kg, m (IQR) 1050 17 (7‐30) 22 (11‐33) 22 (12‐33) 19 (11‐29) 14 (7‐23)
Weight gain >20 kg since age 20, % 44.6 55.4 55.8 48.4 36.5
PR (95% CI) 0.80 (0.71‐0.90) 0.99 (0.92‐1.07) 1 (ref) 0.87 (0.80‐0.94) 0.66 (0.56‐0.77)
Blood pressure, mmHg, m (IQR)
HypertensionII, % 30 31.7 35.7 34.4 35.5 33.3
PR (95% CI) 0.92 (0.80‐1.07) 1.04 (0.93‐1.16) 1 (ref) 1.03 (0.94‐1.14) 0.97 (0.83‐1.14)
Total cholesterol, mmol/L, m (IQR) 2096 4.7 (3.9‐5.3) 4.4 (3.8‐5.2) 4.3 (3.7‐5.1) 4.3 (3.7‐5.0) 4.2 (3.6‐4.9)
Triglycerides, mmol/L, m (IQR) 357 2.1 (1.3‐3.2) 1.8 (1.3‐2.7) 1.6 (1.2‐2.3) 1.5 (1.1‐2.1) 1.4 (1.1‐1.9)
HDL cholesterol, mmol/L, m (IQR) 2085 1.0 (0.9‐1.2) 1.1 (1.0‐1.3) 1.2 (1.0‐1.5) 1.3 (1.1‐1.6) 1.3 (1.1‐1.7)
LDL cholesterol, mmol/L, m (IQR) 184 2.4 (1.9‐3.0) 2.4 (1.8‐3.0) 2.2 (1.7‐2.8) 2.1 (1.6‐2.7) 2.0 (1.6‐2.7)
LDL cholesterol > 2.5 mmol/L, % 49.6 45.6 37.9 33.4 30.9
PR (95% CI) 1.31 (1.17‐1.46) 1.21 (1.10‐1.32) 1 (ref) 0.88 (0.80‐0.97) 0.82 (0.69‐0.96)
DyslipidemiaIII, % 895 90.0 86.0 78.9 72.3 68.0
PR (95% CI) 1.14 (1.10‐1.19) 1.09 (1.05‐1.13) 1 0.92 (0.88‐0.96) 0.86 (0.79‐0.93)
CRP, mg/LIV, m (IQR) 4267 3.2 (1.2‐6.4) 2.5 (1.1‐5.8) 1.9 (0.8‐4.4) 1.8 (0.9‐3.7) 1.6 (0.8‐3.4)
CRP >3.0 mg/L, % 53.0 44.4 37.6 29.4 25.5
PR (95% CI) 1.41 (1.12‐1.78) 1.18 (0.95‐1.47) 1 (ref) 0.78 (0.61‐1.01) 0.68 (0.42‐1.11)
Hospital‐diagnosed retinopathy, % 0 7.2 3.5 4.5 3.9 3.8
PR (95% CI) 1.58 (1.08‐2.31) 0.77 (0.52‐1.12) 1 (ref) 0.86 (0.61‐1.20) 0.85 (0.49‐1.45)
Albumin‐creatinine ratio, mg/g, m (IQR) 610 9.0 (5.0‐22.0) 8.0 (3.0‐22.1) 4.0 (9.0‐21.0) 4.0 (9.0‐22.0) 12 (5.8‐36.0)
≥30, % 21.8 19.4 19.1 20.3 27.5
PR (95% CI) 1.14 (0.93‐1.40) 1.02 (0.86‐1.20) 1 (ref) 1.06 (0.91‐1.23) 1.44 (1.18‐1.76)

Estimates are shown as percentages (%), medians (m) with interquartile ranges (IQR), or prevalence ratios (PR) with 95% confidence Intervals (95% CI).
*Adjusted for time since commencement of glucose‐lowering pharmacological treatment to enrolment in the DD2 cohort.
I
Waist‐to‐hip ratio > 0.85/>0.90 for men/women.
II
Systolic blood pressure > 140 mmHg or diastolic blood pressure > 90 mmHg.
III
LDL cholesterol >2.60 mmol/L, or HDL cholesterol <1.02 mmol/L, or triglyceride >1.7 mmol/L.
IV
Only available for a subset of 840 cohort members.

higher likelihood of receiving any GLD, and a 3.6‐fold higher likelihood prevalence of hypertension among the early‐onset individuals was
of receiving several GLDs, within the first year.27 Two large higher than in a similar‐aged background population, whereas the
cohort studies28,29 in Danish background populations identified an prevalence among the late‐onset individuals was lower than in a
age‐gradient with higher prevalence of hypertension in older similar‐aged background population.
individuals. When comparing these findings with the prevalence Health behaviour is likely to contribute to the observed adverse
found in our study (~30% with hypertension in all age groups), the clinical profile of the early‐onset group. A large cross‐sectional study
BO ET AL. 7 of 9

TABLE 3 Behavioural risk factors among 5115 persons with newly diagnosed type 2 diabetes in 5 groups defined by age at diagnosis
Age at Diagnosis
Missing ≤45 years 46‐55 years 56‐65 years 66‐75 years >75 years
Behavioural Risk Factors n n = 516 n = 1091 n = 1651 n = 1466 n = 391

Physical activityI, days/week, m (IQR) 0 3 (2‐5) 3 (1‐6) 4 (2‐7) 4 (2‐7) 4 (1‐7)

0‐3, % 56.2 54.6 48.1 46.3 49.6

PR (95% CI) 1.17 (1.07‐1.28) 1.14 (1.06‐1.22) 1 (ref) 0.96 (0.89‐1.04) 1.03 (0.92‐1.15)

4‐5, % 21.5 20.0 19.5 17.5 15.6

PR (95% CI) 1.10 (0.91‐1.34) 1.03 (0.88‐1.20) 1 (ref) 0.90 (0.77‐1.04) 0.80 (0.62‐1.03)
6‐7, % 22.3 25.4 32.4 36.3 34.8

PR (95% CI) 0.69 (0.58‐0.82) 0.78 (0.69‐0.89) 1 (ref) 1.12 (1.02‐1.23) 1.07 (0.92‐1.25)
Daily smoking, % 271 23.8 23.7 20.3 12.9 7.9
PR (95% CI) 1.18 (0.98‐1.42) 1.17 (1.01‐1.36) 1 (ref) 0.64 (0.54‐0.75) 0.39 (0.27‐0.56)
High‐risk alcohol intakeII, % 0 3.7 6.1 7.3 8.7 4.6
PR (95% CI) 0.50 (0.31‐0.81) 0.84 (0.63‐1.12) 1 (ref) 1.19 (0.94‐1.51) 0.63 (0.39‐1.02)

Estimates are shown as percentages (%), medians (m) with interquartile ranges (IQR), or prevalence ratios (PR) with 95% confidence Intervals (95% CI).
I
Days per week where cohort members report at least 30 minutes of moderate to vigorous physical activity.
II
Alcohol intake >14/21 drinks per week for women/men.

TABLE 4 Therapy among 5115 persons with newly diagnosed type 2 diabetes in 5 groups defined by age at diagnosis

Age at Diagnosis
Missing ≤45 years 46‐55 years 56‐65 years 66‐75 years >75 years
Therapy n n = 516 n = 1091 n = 1651 n = 1466 n = 391

Glucose‐Lowering Drugs 0
No glucose‐lowering drugs, % 8.1 12.3 14.6 17.5 21.5

PR (95% CI) 0.56 (0.41‐0.76) 0.84 (0.69‐1.02) 1 (ref) 1.20 (1.02‐1.41) 1.47 (1.18‐1.84)

Non‐insulin drugs, % 80.0 78.2 78.4 77.7 75.2

PR (95% CI) 1.02 (0.97‐1.07) 1.00 (0.96‐1.04) 1 (ref) 0.99 (0.96‐1.03) 0.96 (0.90‐1.02)

Insulin only, % 1.4 1.8 1.2 0.6 1.0

PR (95% CI) 1.18 (0.50‐2.79) 1.59 (0.85‐2.97) 1 (ref) 0.53 (0.24‐1.18) 0.89 (0.30‐2.60)

Insulin and non‐insulin drugs, % 10.5 7.7 5.9 4.2 2.3

PR (95% CI) 1.78 (1.30‐2.45) 1.31 (0.99‐1.74) 1 (ref) 0.72 (0.53‐0.98) 0.39 (0.20‐0.77)
Duration of glucose‐lowering treatment 12.8 (3.0‐27.5) 14.3 (2.8‐28.8) 17.1 (3.9‐32.1) 18.7 (5.3‐33.8) 21.2 (8.7‐34.5)
before DD2 enrolmentI, month, m(IQR)
Antihypertensive drugs, % 0 39.5 63.4 76.3 83.4 90.0
PR (95% CI) 0.52 (0.46‐0.58) 0.83 (0.79‐0.88) 1 (ref) 1.09 (1.05‐1.13) 1.18 (1.13‐1.23)
Lipid‐lowering drugs, % 0 53.9 67.5 76.6 78.4 74.7
PR (95% CI) 0.70 (0.65‐0.77) 0.88 (0.84‐0.93) 1 (ref) 1.02 (0.99‐1.06) 0.98 (0.92‐1.04)
Anticoagulation drugs, % 0 7.4 21.4 32.3 40.2 44.8
PR (95% CI) 0.23 (0.17‐0.31) 0.66 (0.58‐0.76) 1 (ref) 1.25 (1.33‐1.37) 1.39 (1.22‐1.58)
Eye screening completedII, % 0 56.2 54.5 56.2 52.8 45.0
PR (95% CI) 1.00 (0.92‐1.09) 0.97 (0.90‐1.04) 1 (ref) 0.94 (0.88‐1.00) 0.80 (0.71‐0.90)
Foot screening completedII, % 0 82.0 84.0 86.3 85.9 87.0
PR (95% CI) 0.95 (0.91‐0.99) 0.97 (0.94‐1.01) 1 (ref) 1.00 (0.97‐1.02) 1.01 (0.97‐1.05)

Estimates are shown as percentage (%), median (m) with interquartile range (IQR), or prevalence ratio (PR) with 95% confidence Intervals (95% CI).
I
Months since commencement of glucose‐lowering pharmacological treatment to DD2 enrolment.
II
Examination of foot or eye registered in the year prior to or after enrolment.

found a lower physical activity level among persons with early‐onset not only higher than in the later‐onset groups but is also higher
type 2 DM than among those with later‐onset type 2 DM.16 The than the Danish national average of 18% in a similar age group
smoking prevalence of 24% in the early‐onset group in our study was (35‐44 years)30. Similarly, in the United States, smoking prevalence
8 of 9 BO ET AL.

among less than 40‐year‐old individuals was reported at 25% and 20% condition. Our results underline the need for clinical awareness and
among persons with and without type 2 DM.30 multifactorial interventions among early‐onset type 2 diabetes
As in other studies,16,31 we found that antihypertensive, lipid‐ patients and a need for prospective studies exploring the association
lowering, and anticoagulation drugs were used less often by persons between early risk factors and development of diabetes‐related
with early‐onset type 2 DM, in spite of their worse risk factor profile. complications.
This indicates a problem of under‐treatment. The identified presence
of GAD antibody positive individuals among people with clinically ACKNOWLEDGEMENTS
diagnosed type 2 DM suggests that there may be pitfalls in diabetes We are grateful to all participants of the DD2 cohort.
diagnosing in routine clinical practice.
Our findings are disquieting considering the consistent findings of DISC LOSURE OF INTERESTS
a high risk of later diabetes complications in persons with early‐onset
None declared
type 2 DM.7,9,11,31,32 One study found that after 20 years of diabetes
duration, 37% of persons diagnosed before the age of 40 years had
AUTHOR CONTRIBUTIONS
developed cardiovascular disease, and the early‐onset individuals
HBN, HTS, JR, TKH, JS, SF, TL, JSN, and RWT participated in designing
appeared to develop microvascular complications 13 to 20 years
earlier than later‐onset individuals. 11
Gregg et al. 12
showed that the the DD2 cohort. JSN, RWT, HTM, SKN, and AB conceived the study.

overall improvements in diabetes outcomes observed during the last SKN performed the statistical analysis. AB initially drafted the article,
with help by HTM, RWT, and JSN. All authors contributed substan-
20 years are primarily due to a reduction in complications among older
tially, revised the manuscript for intellectual content, and approved
persons with type 2 DM. Therefore, the increased incidence of type 2
DM in high‐risk young persons could cause a future rise in diabetes the final version to be submitted.

complications.12
STATEMENT OF ALL FUNDING S OURC ES

The Danish Centre for Strategic Research in Type 2 Diabetes is

4.1 | Strength and limitations supported by the Danish Agency for Science (grant nos. 09‐067009

The main strength of our study was the comprehensiveness of uni- and 09‐075724). DD2 is also supported by the Danish Health and

formly collected data in a large incident cohort of persons clinically Medicines Authority, the Danish Diabetes Association, and an unre-

diagnosed with type 2 DM. The linkage of clinical and self‐reported stricted donation from Novo Nordisk A/S. The partners of the project

information with data from high‐quality national health registers are listed on the project website http://www.dd2.nu/. Moreover, this

allowed for a full description of risk factors. For some risk factors, such work was funded by research grants from the Danish Diabetes

as anthropometric measures and laboratory values, there were missing Academy, supported by the Novo Nordisk Foundation, and by the

data, and missing data were somewhat more common in the elderly Central Denmark Region, the Riisfort Foundation, and Aarhus

than younger age groups in our cohort (data not shown). A slightly University. All funding was given as nonrestricted.

lower completeness of risk factor values by older age would not

necessarily bias our findings, unless completeness was related both ORCID

to age and to the actual value of the data, which we find less likely. A. Bo http://orcid.org/0000-0001-9296-7179
Availability of laboratory values was related to calendar period of data R.W. Thomsen http://orcid.org/0000-0001-9135-3474
in the biobank, not to age group. J.S. Nielsen http://orcid.org/0000-0003-3179-1186
The cross‐sectional study design is an inborn limitation, as it H. Beck‐Nielsen http://orcid.org/0000-0002-7428-7898
implies uncertainty regarding how the risk factors preceded each J. Rungby http://orcid.org/0000-0002-5207-623X
other, and impedes knowledge about the development over time in H.T. Sørensen http://orcid.org/0000-0003-4299-7040
risk profiles. Moreover, there is an over‐representation of persons with T.K. Hansen http://orcid.org/0000-0002-0708-2868
type 2 DM receiving hospital‐based versus primary care in the DD2 J. Søndergaard http://orcid.org/0000-0002-1629-1864
cohort. Consequently, since both young age and a high risk‐factor level T. Lauritzen http://orcid.org/0000-0002-4579-5164
in type 2 DM may lead to referral from GP to outpatient hospital care, H.T. Maindal http://orcid.org/0000-0003-0525-7254
relatively more high‐risk individuals may have been recruited into the
cohort among early‐ than among later‐onset type 2 DM individuals, RE FE RE NC ES
leading to a Berkson‐like bias and a possible overestimation of risk 1. Alberti G, Zimmet P, Shaw J, et al. Type 2 diabetes in the young: the
factor prevalence in the early‐onset group. evolving epidemic: the international diabetes federation consensus
workshop. Diabetes Care. 2004;27(7):1798‐1811.

2. Song SH, Hardisty CA. Early‐onset type 2 diabetes mellitus: an

5 | C O N CL U S I O N increasing phenomenon of elevated cardiovascular risk. Expert Rev
Cardiovasc Ther. 2008;6(3):315‐322.
In conclusion, our study identified an increasing prevalence of
3. Wilmot EG, Davies MJ, Yates T, Benhalima K, Lawrence IG, Khunti K.
clinical and behavioural risk factors the younger the onset age of Type 2 diabetes in younger adults: the emerging UK epidemic. Postgrad
type 2 DM, emphasizing that early‐onset type 2 DM is not a benign Med J. 2010;86(1022):711‐718.
BO ET AL. 9 of 9

4. Holden SH, Barnett AH, Peters JR, et al. The incidence of type 2 21. Nielsen JS, Thomsen RW, Steffensen C, Christiansen JS. The
diabetes in the United Kingdom from 1991 to 2010. Diabetes Obes Danish Centre for Strategic Research in Type 2 Diabetes (DD2) study:
Metab. 2013;15(9):844‐852. implementation of a nationwide patient enrollment system. Clin
5. Lammi N, Taskinen O, Moltchanova E, et al. A high incidence of type 1 Epidemiol. 2012;4(Suppl 1):27‐36.
diabetes and an alarming increase in the incidence of type 2 diabetes 22. Thomsen RW, Friborg S, Nielsen JS, Schroll H, Johnsen SP. The
among young adults in Finland between 1992 and 1996. Diabetologia. Danish Centre for Strategic Research in Type 2 Diabetes (DD2): organi-
2007;50(7):1393‐1400. zation of diabetes care in Denmark and supplementary data sources for
6. Li L, Ji L, Guo X, et al. Prevalence of microvascular diseases among data collection among DD2 study participants. Clin Epidemiol.
tertiary care Chinese with early versus late onset of type 2 diabetes. 2012;4(Suppl 1):15‐19.
J Diabetes Complications. 2015;29(1):32‐37. 23. Dansk Selskab for Almen Medicin. Guidelines for type 2 diabetes.
7. Hillier TA, Pedula KL. Complications in young adults with early‐onset 2011. Available at http://www.irf.dk/download/Publikationer/
type 2 diabetes: losing the relative protection of youth. Diabetes Care. vejledninger/diabetesfolder.pdf Accessed 17 March 2017
2003;26(11):2999‐3005. 24. American Diabetes Association. Management of dyslipidemia in adults
8. Chan JC, Lau ES, Luk AO, et al. Premature mortality and comorbidities with diabetes. Diabetes Care. 2003;26(suppl 1)
in young‐onset diabetes: a 7‐year prospective analysis. Am J Med. 25. Benhalima K, Wilmot E, Khunti K, Gray LJ, Lawrence I, Davies M. Type
2014;127(7):616‐624. 2 diabetes in younger adults: clinical characteristics, diabetes‐related
9. Huo X, Gao L, Guo L, et al. Risk of non‐fatal cardiovascular diseases in complications and management of risk factors. Prim Care Diabetes.
early‐onset versus late‐onset type 2 diabetes in China: a cross‐sectional 2011;5(1):57‐62.
study. Lancet Diabetes Endocrinol. 2016;4(2):115‐124. 26. Lind M, Pivodic A, Cea‐Soriano L, Nerman O, Pehrsson NG,
10. Song SH, Gray TA. Early‐onset type 2 diabetes: high risk for premature Garcia‐Rodriguez LA. Changes in HbA1c and frequency of measuring
diabetic retinopathy. Diabetes Res Clin Pract. 2011;94(2):207‐211. HbA1c and adjusting glucose‐lowering medications in the 10 years
following diagnosis of type 2 diabetes: a population‐based study in
11. Song SH, Hardisty CA. Early onset type 2 diabetes mellitus: a harbinger
the UK. Diabetologia. 2014;57(8):1586‐1594.
for complications in later years ‐ clinical observation from a secondary
care cohort. QJM. 2009;102(11):799‐806. 27. Mor A, Berencsi K, Svensson E, et al. Prescribing practices and clinical
predictors of glucose‐lowering therapy within the first year in
12. Gregg EW, Sattar N, Ali MK. The changing face of diabetes complica-
people with newly diagnosed type 2 diabetes. Diabet Med.
tions. Lancet Diabetes Endocrinol. 2016;4(6):537‐547.
2015;32(12):1546‐1554. Epub 2015 Jul 3
13. Svensson E, Berencsi K, Sander S, et al. Association of parental history
28. Sehestedt T, Ibsen H, Jørgensen T. Awareness, treatment and control
of type 2 diabetes with age, lifestyle, anthropometric factors, and
of hypertension in Denmark. The Inter99 study. Blood Press.
clinical severity at type 2 diabetes diagnosis: results from the DD2
2007;16(5):312‐319.
study. Diabetes Metab Res Rev. 2016;32(3):308‐315.
29. Christensen AI, Ekholm O, Davidsen M, Juel K. Sundhed og sygelighed I
14. Molyneaux L, Constantino M, Yue D. Strong family history predicts
Danmark, og udvikling siden 1987. Statens Institut for Folkesundhed,
a younger age of onset for subjects diagnosed with type 2 diabetes.
SDU. Available at http://www.si-folkesundhed.dk/upload/susy_2010_
Diabetes Obes Metab. 2004;6(3):187‐194.
3_2_specifikke_sygdomme_og_lidelser.pdfAccessed September 2017
15. Gunathilake W, Song S, Sridharan S, Fernando DJ, Idris I. Cardiovascu-
30. Fan AZ, Rock V, Zhang X, Li Y, Elam‐Evans L, Balluz L. Trends in cigarette
lar and metabolic risk profiles in young and old patients with type 2
smoking rates and quit attempts among adults with and without diagnosed
diabetes. QJM. 2010;103(11):881‐884.
diabetes, United States, 2001‐2010. Prev Chronic Dis. 2013;10:E160
16. Yeung RO, Zhang Y, Luk A, et al. Metabolic profiles and treatment
31. Song SH, Gray TA. Early intensive cardiovascular risk management in
gaps in young‐onset type 2 diabetes in Asia (the JADE programme): a
young people with type 2 diabetes. Diabetes Res Clin Pract.
cross‐sectional study of a prospective cohort. Lancet Diabetes
2011;92(3):e70‐e72.
Endocrinol. 2014;2(12):935‐943.
32. Wong J, Molyneaux L, Constantino M, Twigg SM, Yue DK. Timing is
17. Hatunic M, Burns N, Finucane F, Mannion C, Nolan JJ. Contrasting
everything: age of onset influences long‐term retinopathy risk in type
clinical and cardiovascular risk status between early and later onset
2 diabetes, independent of traditional risk factors. Diabetes Care.
type 2 diabetes. Diab Vasc Dis Res. 2005;2(2):73‐75.
2008;31(10):1985‐1990.
18. Hillier TA, Pedula KL. Characteristics of an adult population with newly
diagnosed type 2 diabetes: the relation of obesity and age of onset.
Diabetes Care. 2001;24(9):1522‐1527. How to cite this article: Bo A, Thomsen RW, Nielsen JS, et al.
19. Aguilar‐Salinas CA, Rojas R, Gomez‐Perez FJ, et al. Prevalence and Early‐onset type 2 diabetes: Age gradient in clinical and behav-
characteristics of early‐onset type 2 diabetes in Mexico. Am J Med.
ioural risk factors in 5115 persons with newly diagnosed type 2
2002 Nov;113(7):569‐574.
diabetes—Results from the DD2 study. Diabetes Metab Res Rev.
20. Smith TL, Drum ML, Miernik J, Fogelfeld LA, Lipton RB. Early and later
onset type 2 diabetes in uninsured patients: clinical and behavioral dif- 2018;34:e2968. https://doi.org/10.1002/dmrr.2968
ferences. J Health Care Poor Underserved. 2008;19(4):1119‐1135.