Main topic

English version of “Prävalenz und zeitliche C. Heidemann1 · Y. Du1 · I. Schubert2 · W. Rathmann3 · C. Scheidt-Nave1 Entwicklung des bekannten Diabetes mellitus. 1 Department of Epidemiology and Health Monitoring, Robert Koch Institute, Ergebnisse der Studie zur Gesundheit 2 PMV Research Group, Clinic and Polyclinic for Childhood and Adolescent Erwachsener in Deutschland (DEGS1)” Bundesgesundheitsbl 2013 · 56:668–677 Psychiatry and Psychotherapy, University of Cologne DOI 10.1007/s00103-012-1662-5 3 German Diabetes Centre, Institute for Biometrics and Epidemiology, Leibniz © Springer-Verlag Berlin Heidelberg 2013 Centre for Diabetes Research at Heinrich Heine University, Düsseldorf Prevalence and temporal trend of known diabetes mellitus Results of the German Health Interview and Examination Survey for Adults (DEGS1)

Background and purpose trolled diagnosed diabetes leads to dam- On the basis of data from the Health age of the blood vessels and peripheral Interview and Examination Survey con- Diabetes mellitus is a chronic metabol- nerves. This results in an increased risk of ducted from 2008 to 2011, the following ic disease, which is characterised by in- cardiovascular diseases such as heart at- article presents up-to-date, representative creased blood sugar concentrations as a tack and stroke, as well as of kidney fail- estimates regarding lifetime prevalence of result of an absolute or relative lack of in- ure, blindness and foot amputations [1, 3]. known diabetes in the resident population sulin. An absolute lack of insulin charac- These health complications, in addition to of aged 18–79 years. The tempo- terises “type-1 diabetes”, which manifests the diabetes itself, lead to a lesser quality ral trend is shown in comparison with the itself mainly during childhood or adoles- of life and reduced life expectancy of those last Health Interview and Examination cence. It is caused by an autoimmune de- affected [3, 4], as well as to high costs for Survey from 1997 to 1999. The results are struction of the insulin-producing beta the health care system [5]. presented and discussed in the context of cells of the pancreas [1]. A relative lack International analyses show that the previous studies in Germany. For the pur- of insulin characterises “type-2 diabetes”, number of adults with diabetes world- pose of assessing the prevalence and tem- which primarily occurs in adulthood. wide has more than doubled in the last poral trend of undiagnosed diabetes, labo- Here, the impaired glucose metabolism three decades [6]. This trend mainly re- ratory data that are comparable over time results from a reduced insulin effect in flects the increase in type-2 diabetes [2]. are required. This is currently being cross- conjunction with inadequately compen- It is assumed that the disease has not yet calibrated for both surveys and is there- sating insulin secretion. In addition to been diagnosed in approximately half of fore not included as part of the paper. genetic predisposition, the main risk fac- those affected [7]. tors include a diet related to the “West- German studies on prevalence esti- Methods ern” lifestyle, lack of exercise and associ- mates of known diabetes are based most- ated excess weight [1, 2]. A relative lack of ly on data from regional studies, health Study design insulin can also develop for the first time insurance companies or general practices during pregnancy, which generally sub- assessed at least 5–10 years ago (see sum- The “German Health Interview and Ex- sequently disappears (“gestational diabe- mary in . Tab. 6). Nationwide, popula- amination Survey for Adults” (DEGS) tes”). However, a high risk remains of de- tion-based data on known diabetes were is part of the health monitoring sys- veloping a manifest type-2 diabetes in lat- provided most recently via the telephone tem at the Robert Koch Institute (RKI). er life [1]. surveys “German Health Update” (GE- The concept and design of DEGS are de- Long-term increased blood sugar con- DA) 2009 [8, 9] and GEDA 2010 [10]. scribed in detail elsewhere [12, 13, 14, 15, centrations in the case of as-yet undiag- There are only isolated studies on undiag- 16]. The first wave of the survey (DEGS1) nosed diabetes or an inadequately con- nosed diabetes [11]. was conducted from 2008 to 2011 and

Bundesgesundheitsblatt - Gesundheitsforschung - Gesundheitsschutz 5/6 · 2013 | 1 Main topic

Tab. 1 Lifetime prevalence (percent, 95% confidence interval) of known diabetes according One participant that stated“diagnosis dur- to gender and age groups. Nunweighted=7,080 ing pregnancy” was pregnant when inter- Age group in 18–39 40–49 50–59 60–69 70–79 Overall viewed. Because of additional informa- years tion (diagnosis 3 years ago, diabetes exist- Gender ing in the past 12 months, current treat- Female 3.7 4.5 4.0 10.7 21.8 7.4 ment with insulin and treatment immedi- (2.5–5.5) (3.0–6.8) (2.6–6.0) (8.2–13.8) (17.6–26.7) (6.5–8.5) ately after diagnosis by tablet), type-2 dia- Male 0.9 2.0 7.3 17.0 22.0 7.0 betes rather than gestational diabetes was (0.3–2.3) (1.1–3.7) (5.3–10.1) (13.1–21.7) (17.6–27.2) (6.0–8.1) assumed in this case. Overall 2.3 3.2 5.7 13.8 21.9 7.2 (1.5–3.4) (2.3–4.6) (4.4–7.2) (11.4–16.6) (18.7–25.5) (6.5–8.0) Statistical analysis

comprised interviews, examinations and Definition of known diabetes The cross-sectional and trend analyses tests [17, 18]. The target population com- were conducted using a weighting factor, prises the residents of Germany aged Lifetime prevalence of known diabetes which corrects sample deviations from the 18–79 years. DEGS1 has a mixed design, was established using the following case population structure (as of 31 December which allows for both cross-sectional definition: 2010) with regard to age, sex, region and and longitudinal analyses. For this pur- F Answering yes to the question “Have nationality, as well as type of communi- pose, a random sample from local pop- you ever been diagnosed with dia- ty and education [16]. A separate weight- ulation registries was drawn to complete betes by a doctor?” as part of a stan- ing factor was created for the examina- the sample of participants of the “Ger- dardised, computer-aided and physi- tion part. Calculation of the weighting man National Health Interview and Ex- cian-administered interview or factor also considered the re-participa- amination Survey 1998” (GNHIES98) F Taking of anti-diabetics having been tion probability of the GNHIES98 partic- who re-participated. A total of 8,152 per- documented as part of an automated ipants based on a logistic regression mod- sons participated, including 4,193 first- assessment of medication taken with- el. For the purpose of conducting trend time invitees (response rate: 42%) and in the past 7 days. analyses, the data from GNHIES98 were 3,959 revisiting participants of the GN- age-adjusted to the population structure HIES98 (response rate: 62%). In all, The physician-administered interview as of 31 December 2010. A non-response 7,238 persons visited one of the 180 ex- did not include any direct questions on analysis and a comparison of selected in- amination centres, and 914 were inter- the type of diabetes. However, the pro- dicators with data from census statistics viewed only. The net sample allows for portion of type-1 diabetes was estimated indicated a high level of sample repre- representative cross-sectional and trend by gathering the following additional in- sentativeness for the resident population analyses for the age range of 18–79 years formation: of Germany aged 18–79 years [16]. In or- (n=7,988, including 7,116 in examination F Age at diagnosis <30 years, and der to take into account both the weight- centres) in comparison with GNHIES98 F Insulin treatment immediately after ing and the correlation of the participants ((n=7,124) [16]. The data of the revisit- diagnosis, and within a community, the confidence inter- ing participants are suitable for longitu- F Current insulin treatment. vals and p values were determined using dinal analyses. the survey procedures for complex sam- The cross-sectional analyses on In addition, women with a diagnosis of ples in SAS 9.2. Differences with p values known diabetes are based on the da- diabetes before the age of 50 years were of <0.05 were considered to be statistical- ta of 7,116 participants aged 18–79 years asked in the DEGS1 whether they were ly significant. who completed the examination part of pregnant at the time of diagnosis. The Estimates of the lifetime prevalence of DEGS1. In an additional analysis, the proportion of gestational diabetes was es- diabetes in cross-sectional and trend anal- diabetes prevalence established in this timated by linking with additional infor- yses are presented for the population aged sample was compared with that in the mation as follows: 18–79 years in general, as well as stratified total net sample (n=7,988), which also F Diagnosis during pregnancy, and by gender. Additional stratification vari- contained data from those surveyed by F Not currently taking anti-diabetics, ables in the cross-sectional analyses are interview only. The participants of GN- and age, socioeconomic status, residential re- HIES98 (n=7,124) were included in the F No existing diabetes in the past gion and health insurance company. So- temporal comparison. Participants with 12 months (n=32) or in the case of cioeconomic status was determined us- missing data regarding diabetes diagno- missing data (n=6) or in the case of ing an index that includes information re- sis were excluded from the analyses, pro- unclear answer (n=4) with regard garding school education and vocation- vided they were not taking anti-diabetics to diabetes existing within the past al training, professional status and net (DEGS1: n=36; GNHIES98: n=25). 12 months: HbA1c <6.5% and glucose household income (weighted by house- <7.0 mmol/l (fasting)/<11.1 mmol/l hold needs) and which allows classifica- (non-fasting). tion into either the low, middle or high

2 | Bundesgesundheitsblatt - Gesundheitsforschung - Gesundheitsschutz 5/6 · 2013 Abstract · Zusammenfassung status group [19]. The residential region Bundesgesundheitsbl 2013 · DOI 10.1007/s00103-012-1662-5 was stratified into new federal states (in- © Springer-Verlag Berlin Heidelberg 2013 cluding Berlin) and old federal states for C. Heidemann · Y. Du · I. Schubert · W. Rathmann · C. Scheidt-Nave the purpose of comparisons with previous Prevalence and temporal trend of known diabetes mellitus. surveys. For further regional differentia- Results of the German Health Interview and tion, the old federal states were sub-divid- Examination Survey for Adults (DEGS1) ed into North (Schleswig-Holstein, Ham- burg, Lower , ), Central Abstract (North Rhine-Westphalia, , Rhine- The first wave of the “German Health Inter- cioeconomic status. Prevalence varied de- land Palatinate, ) and South (Ba- view and Examination Survey for Adults” pending on the type of health insurance held (DEGS1, 2008–2011) allows for up-to-date, and was highest amongst those insured with varia, Baden-Wuerttemberg). Health in- representative prevalence estimates of AOK health insurance funds. In comparison surance companies were categorised into known diabetes amongst the 18- to 79-year- with GNHIES98, there was a 38% increase in statutory health insurance, private health old resident population of Germany. Tempo- prevalence, of which approximately one third insurance as comprehensive health insur- ral trends can be shown by comparing the is to be attributed to demographic ageing. In ers and other health care provision (bene- survey findings with those of the “German the context of other nationwide studies, the fit aid, foreign health insurance, other en- National Health Interview and Examination results indicate a figure of at least 4.6 million Survey 1998” (GNHIES98). The definition of 18- to 79-year-olds having been diagnosed titlement, no health insurance/self-payer). known diabetes was based on self-reports in with diabetes at some point. Planned analy- In addition and analogous to earlier tele- physician-administered interviews that asked ses of undiagnosed diabetes will contribute phone surveys, the statutory health insur- respondents if they had ever been diagnosed to the interpretation of the observed increase ance providers were sub-divided into lo- with diabetes by a doctor or were on anti-dia- in the prevalence of known diabetes. cal health insurance funds (AOK), sub- betic medication. Overall, diabetes had been stitute health insurance funds, company diagnosed in 7.2% of the adults (7.4% of the Keywords women; 7.0% of the men). The prevalence in- Diabetes mellitus · Germany · Adults · Health health insurance funds and other statu- creased substantially with advancing age and survey · Prevalence tory health insurance funds. By compar- was higher in persons of low than of high so- ing non-adjusted and age-adjusted logis- tical regression models, the influence of age on the relationship between the re- Prävalenz und zeitliche Entwicklung des bekannten spective stratification variable and preva- Diabetes mellitus. Ergebnisse der Studie zur lence of diabetes was evaluated. Gesundheit Erwachsener in Deutschland (DEGS1) Zusammenfassung Results Die erste Welle der „Studie zur Gesundheit iert je nach Krankenversicherungsart und ist ­Erwachsener in Deutschland“ (DEGS1, 2008– für Versicherte der Allgemeinen Ortskran- Prevalence of known diabetes 2011) ermöglicht aktuelle, repräsentative kenkasse (AOK) am höchsten. Im Vergleich Prävalenzschätzungen des bekannten­ Di- zum BGS98 zeigt sich ein Prävalenzanstieg In total, 591 of the 7,080 participants of the abetes in der 18- bis 79-jährigen Wohnbe­ um 38%, wovon etwa ein Drittel auf die de- völkerung. Im Vergleich zum „Bundes-Ge- mografische Alterung zurückzuführen ist. DEGS1 examination part stated they had sundheitssurvey 1998“ (BGS98) lässt sich die Die Ergebnisse weisen im Kontext mit an- been diagnosed with diabetes by a doctor zeitliche Entwicklung darstellen. Die Defi- deren bundesweiten Studien auf aktuell min­ at some point. . Tab. 1 shows the corre- nition des bekannten Diabetes beruht auf destens 4,6 Mio. 18- bis 79-Jährige mit einer ­ sponding lifetime prevalence of known Selbstangaben zu einem jemals ärztlich fest­ Diabetesdiagnose hin. Geplante Analysen diabetes for the population aged 18– gestellten Diabetes in ärztlichen Interviews zum nicht diagnostizierten Diabetes werden 79 years of 7.2% overall (women: 7.4%; oder der Einnahme von Antidiabetika. Ins­ zur Interpretation des beobachteten Präva- gesamt wurde bei 7,2% der Erwachsenen lenzanstiegs des bekannten Diabetes bei­ men: 7.0%). For both sexes the prevalence (7,4% der Frauen; 7,0% der Männer) jemals tragen. increases substantially with age from un- ein Diabetes diagnostiziert. Die Prävalenz der 5% amongst those under 50 years of steigt mit zunehmendem Alter deutlich an Schlüsselwörter age to around 22% in the 70- to 79-year- und ist bei niedrigem Sozialstatus höher als Diabetes mellitus · Deutschland · olds. bei hohem Sozialstatus. Die Prävalenz vari- Erwachsene · Gesundheitssurvey · Prävalenz An analysis of the total net sample, which also included participants surveyed by means of interview only, yielded com- 1.1% of the total prevalence of known di- tes for women and men of low socioeco- parable results with 677 of the 7,934 par- abetes. The prevalence of gestational dia- nomic status than for those of high socio- ticipants included having a diabetes diag- betes amongst women amounted to 1.2% economic status. This difference is more nosis or a lifetime prevalence overall of (n=42). This corresponds to a proportion strongly pronounced in women than in 7.4% (women: 7.5%; men: 7.2%). of 16.3% of the total prevalence of known men. While women of middle socio- The prevalence of type-1 diabetes was diabetes amongst women. economic status also show a significant- ascertained as being around 0.1% (n=8). . Tab. 2 reflects a significantly high- ly higher prevalence than those of high This corresponds to a proportion of er lifetime prevalence of known diabe- socioeconomic status, there is no differ-

Bundesgesundheitsblatt - Gesundheitsforschung - Gesundheitsschutz 5/6 · 2013 | 3 Main topic

Tab. 2 Lifetime prevalence (percent, 95% confidence interval) of known diabetes according a total of 3.3 million of 18- to 79-year- to gender and socioeconomic status. Nunweighted=7,012 olds living in Germany had at some point Socioeconomic status Low Middle High been diagnosed with diabetes. By con- Gender trast, according to the estimations based Female 11.6 (8.6–15.5) 7.4 (6.3–8.7) 3.0 (2.0–4.5) on DEGS1 data, diabetes has been diag- Male 10.1 (7.5–13.5) 6.1 (5.1–7.4) 6.2 (4.6–8.3) nosed at some point in a total of 4.6 mil- lion people in this age group. Overall 10.9 (8.8–13.5) 6.8 (6.0–7.7) 4.8 (3.7–6.0) Socioeconomic status: index that includes information on school and vocational education, professional position and net household income (weighted by household needs) Discussion

Tab. 3 Lifetime prevalence (percent, 95% confidence interval) of known diabetes according According to the current DEGS1 data, to gender and residential area. Nunweighted=7,080 7.2% or 4.6 million of the 18- to 79-year- Residential Old federal states New federal old population of Germany have at some area states point been diagnosed with diabetes. The Overall North Central South lifetime prevalence of known diabetes in- creases substantially with age and is high- Gender er amongst those of low socioeconomic Female 7.1 (6.0–8.4) 7.1 (4.6–10.9) 7.7 (6.3–9.4) 6.3 (4.6–8.6) 8.7 (6.7–11.3) status than high socioeconomic status. Male 6.9 (5.7–8.2) 6.3 (4.3–9.3) 7.5 (5.6–10.0) 6.3 (4.7–8.4) 7.5 (5.9–9.4) Prevalence is higher in persons insured in Overall 7.0 (6.2–7.9) 6.7 (5.2–8.7) 7.6 (6.3–9.2) 6.3 (5.1–7.8) 8.1 (6.9–9.5) statutory health insurance schemes—es- Old federal states: North: Schleswig-Holstein, , , Bremen; Central: North Rhine-Westpha- lia, Hessen, Rhineland Palatinate, Saarland; South: , Baden-Wuerttemberg pecially those insured with AOK—than it is amongst those with private medical in- surance. In addition, there are indications ence in prevalence between men of mid- various statutory health insurance funds of regional differences, especially with re- dle and high socioeconomic status. These amongst women and between the statu- gard to a tendency toward a higher preva- results were also observed in the age-ad- tory and the private funds amongst men. lence in the new federal states compared justed model. This pattern is also observed in the age- to the old ones. In comparison to GN- In the stratification of the lifetime adjusted model, although the differences HIES98, the prevalence of known diabe- prevalence of known diabetes according are less pronounced. tes has increased by 38%. An increase of to residential area, as shown in . Tab. 3, 14% is to be attributed to the demograph- there tends to be a higher prevalence for Trends in known diabetes ic ageing of the population. the new federal states than for the old fed- eral states, which is more pronounced In . Tab. 5, the lifetime prevalence of Prevalence of known diabetes amongst women than men. Further strat- known diabetes from the earlier exam- ification of the old federal states into ination survey is compared with that The most recent health surveys conduct- North, Central and South indicates that from DEGS1. The lifetime prevalence ob- ed at population level as telephone in- there tends to be a higher prevalence in served on the basis of the GNHIES98 data terviews, GEDA 2009 and GEDA 2010, the Central region compared to the North (weighted to 1997 population structure) yielded a pooled lifetime prevalence of or South. These observed trends, howev- was 5.2% overall (women: 5.7%; men: 8.7% or a figure of 5.9 million adults aged er, are not statistically significant. They 4.7%). This resulted in a significant ab- 18 years and above having been diag- remain in the age-adjusted model but are solute increase in prevalence over time nosed with diabetes at some point. Lim- even less pronounced. of 2.0% (women: 1.7%; men: 2.3%) and iting the GEDA data to the age range . Tab. 4 shows differences in the life- a relative increase of 38% (women: 30%; of 18–79 years results in a prevalence of time prevalence of known diabetes de- men: 49%). A significant increase also ex- 8.2%, which exceeds the observed preva- pending on the type of health insurance. ists after taking the demographic ageing lence of 7.2% in DEGS1. Conversely, a di- Statutory health insurees have a higher of the population into account (weighted abetes prevalence in the age group over prevalence overall (7.5%) than those pri- to 2010 population structure). This is 1.4% 80 years observed in GEDA and oth- vately insured (3.8%). Amongst persons absolute (women: 1.3%; men: 1.6%) and er studies, similar in size to that in the insured in statutory schemes, those cov- 24% relative (women: 21%; men: 30%). 70–79-year age range [20, 21, 22, 23, 24, ered through AOK have the highest prev- Consequently, an absolute increase of 25, 26], results in a figure of only 5.5 mil- alence at 9.0%, whereas those covered 0.6% (women: 0.4%; men: 0.7%) and rel- lion adults over 18 years with known dia- through company health insurance funds ative increase of 14% (women: 9%; men: betes when the data are transposed onto have the lowest prevalence at 5.9%. This 19%) can be attributed to demographic DEGS1. This difference in size of the esti- pattern is also evident in the gender-strat- ageing. mates between DEGS1 and GEDA is pre- ified analysis, although the prevalence Extrapolations of the GNHIES98 da- sumably a result of the difference in sam- difference is less pronounced between the ta show that approximately 10 years ago ple composition due to a different sam-

4 | Bundesgesundheitsblatt - Gesundheitsforschung - Gesundheitsschutz 5/6 · 2013 Tab. 4 Lifetime prevalence (percent, 95% confidence interval) of known diabetes according to gender and health insurance fund. ­Nunweighted=6,938 Health insur- Statutory health insurance funds Private health Other health ance fund insurance funds care provision (PKV) Overall Local health Substitute health Company health Other statutory insurance insurance funds insurance funds insurance funds (AOK) (EKK) (BKK) Gender Female 7.8 (6.7–9.0) 8.6 (6.6–11.2) 7.3 (5.7–9.2) 6.8 (4.7–9.9) 8.9 (5.7–13.6) 1.2 (0.3–3.9) 4.7 (2.4–9.3) Male 7.2 (6.1–8.5) 9.5 (7.4–12.0) 6.7 (5.0–8.8) 5.0 (3.2–7.7) 7.0 (4.5–10.5) 5.0 (2.8–8.7) 4.7 (2.6–8.4) Overall 7.5 (6.7–8.4) 9.0 (7.5–10.8) 7.0 (5.9–8.3) 5.9 (4.4–7.8) 7.9 (5.7–10.8) 3.8 (2.2–6.5) 4.7 (3.1–7.0)

Tab. 5 Lifetime prevalence (percent, 95% confidence interval) of known diabetes according to gender compared over time. N(DEGS1)unweighted=7,080, N(GNHIES98)unweighted=7,099 tive overall prevalence of known diabe- Survey GNHIES98a GNHIES98b DEGS1b tes in these samples is between approxi- Gender mately 3% and 6% and is therefore higher Female 5.7 (4.8–6.7) 6.1 (5.1–7.2) 7.4 (6.5–8.5) than the DEGS1 figure of about 1%. Given Male 4.7 (4.1–5.4) 5.4 (4.7–6.3) 7.0 (6.0–8.1) this difference in size, the rather conser- Overall 5.2 (4.6–5.9) 5.8 (5.1–6.5) 7.2 (6.5–8.0) vative estimation of type-1 diabetes prev- aWeighted to population distribution as of 31 December 1997 alence in DEGS1 by using the available in- bWeighted to population distribution as of 31 December 2010 formation, along with variations in study design, has to be taken into account. Da- ta on the prevalence of gestational diabe- pling procedure (local population regis- compared to those otherwise insured. In tes were not reported in any of the stud- try sample vs. private households avail- addition to deviations in age structure, ies summarised in . Tab. 6. able via landline) and a different survey which can largely be taken into account method (personal physician-adminis- in age-standardised analyses, differences Temporal trends in known diabetes tered interview vs. telephone interview). in educational status and in the existence For a direct comparison, the DEGS1 data of co-morbidities play a role here [29]. The first time series for Germany regard- were consequently compared exclusively The results of further studies regard- ing the prevalence of known diabetes is with the GNHIES98 data, which are the ing known diabetes in Germany that based on data from the diabetes register of most recent data gathered by a compara- are shown in . Tab. 6 are not compara- the former GDR, which existed between ble method. ble with the current results because the 1960 and 1989. For this period, a contin- The nationwide surveys conduct- study periods were before the DEGS1 pe- ual increase in prevalence from 0.6% to ed on a postal-written basis by the Ber- riod. Regarding the relatively high prev- 4.1% was observed, which was mostly at- telsmann Healthcare Monitor indicate a alence of known diabetes in the patient tributable to the increase in prevalence in prevalence for known diabetes of 8.0% in samples of general practices, it should be the over 50-year age group [20, 39]. Data the 18–79-year age range for 2008 [27]. noted that more people with diagnosed on the prevalence trend in the old federal This prevalence is therefore between the diabetes are to be expected amongst the states do not exist for this period. lifetime prevalence established in DEGS1 practice patients than in the general pop- For the subsequent period until about (2008–11) at 7.2% and that of GEDA ulation. Overall and similarly to DEGS1, 2000, based on the data from population- (2008–10) at 8.2%. On the other hand, the studies summarised in . Tab. 6—in based studies, there is no evidence of a the AOK Hesse/KV Hesse sample points addition to the aforementioned differ- further increase in prevalence. No trend to a prevalence of 9.8% for insured per- ences depending on the type of health in- in diabetes prevalence was observed ei- sons of all age groups for 2009 [5] (18- to surance—point toward a clear increase in ther in the surveys conducted as part of 79-year-olds, excluding in-patient care prevalence with increasing age, especially the German Cardiovascular Prevention cases: 10.2%), which is higher than the from the age of 50 years onwards [8, 20, Study in the old federal states for 25- to prevalence found in GEDA (2008–2010) 21, 22, 23, 24, 25, 26, 28, 30, 31, 32, 33, 34, 69-year-olds between 1984–1986 and for adults aged 18 years and above at 8.7% 35, 36, 37], as well as social differences [8, 1990–1991 [31] or in the KORA/MON- (18- to 79-year-olds: 8.2%). This expected 27] and regional trends [8, 10, 27, 32] in ICA surveys in the Augsburg region for difference is, however, in agreement with the prevalence of diabetes. 25- to 64-year-olds between 1984–1985 the present observations from DEGS1 and Data regarding the prevalence of and 1999–2001 or for 25- to 74-year-olds other current analyses [27, 28], according type-1 diabetes are only to be found in a between 1989–1990 and 1999–2001 [35]. to which persons insured with AOK have few patient samples [25, 37, 38]. The pro- Similarly, the comparison of data from the highest prevalence of diabetes when portion of type-1 diabetes of the respec- the nationwide ­examination surveys

Bundesgesundheitsblatt - Gesundheitsforschung - Gesundheitsschutz 5/6 · 2013 | 5 Main topic

Tab. 6 Overview of studies in Germany with data on the prevalence of known diabetes in adults Study population Period (Stan- Definition of diabetes Prevalence of diabetes dardisation) Nationwide interview and examination survey DEGS1; Representative survey of resident 2008–2011 Self-report on physician- Overall: 7.2% (women: 7.4%, men: 18–79 years, N=7.080 population in Germany (31/12/2010) diagnosed diabetes or 7.0%) (. Tab. 4, . Tab. 5) anti-diabetic medication Type-1: 0.1%, Gestational diabetes/ women: 1.2% AOK: 9.0%; EKK: 7.0%; BKK: 5.9%; other statutory insurance: 7.9%; PKV: 3.8%; other provision: 4.7% GNHIES98; 1997–1999 Overall: 5.2% (women: 5.7%, men: 18–79 years, N=7.099 (31/12/1997) 4.7%) (. Tab. 5) [30]a East/West Health Survey 91, representative survey of residential 1990–1992 Self-report on diabetes Women: 4.7%, men: 5.0% population in the new and old federal states; (1991) diagnosis 25–69 years, N≈7.450 [30] German Cardiovascular Prevention Study, representative survey 1990/91, 1990/1991: women: 4.1%, men: 4.6% of residential population in the old federal states; 1987/88, 1987/1988: women: 3.8%, men: 5.3% 25–69 years, N≈5.000 (per survey) [31] 1984–1986 1984–1986: women: 4.6%, men: 4.5% (1984) Nationwide telephone survey GEDA 2010; Representative survey of German 2009/2010 Self-report on physician- Overall: 8.6% (women: 8.8%; men: 18+ years, N=22.050 [10] speaking resident population in (31/12/2008) diagnosed diabetes 8.5%) GEDA 2009; Germany available via land-line 2008/2009 Overall: 8.8% (women: 9.3%, men: 18+ years, N=21.262 [8, 9] telephone (31/12/2007) 8.2%) GSTel03; 2002/2003 Overall: 6.1% (women: 6.8%, men: 18+ years, N=8.318 [8, 9] (31/12/2001) 5.4%) GSTel03, GSTel04 (pooled); 2002–2004 Overall: 6.5% 18+ years, N=15.354 [28] (31/12/2001) AOK: 10.1%; EKK: 5.6%; BKK: 4.7%; other statutory insurance: 6.6%; PKV: 4.8% Nationwide postal survey Bertelsmann Healthcare Monitor, representative survey of Ger- 2004–2008 Self-report on diabetes Overall 6.9% (women: 5.8%, men: man-speaking resident population in Germany; diagnosis (≥1 physician 8.0%) 18–79 years, N=15.089 (N≈1.500 per survey) [27] visit per quarter or regu- 2008: 8.0%, 2007: 7.2%, 2006: 6.5%, lar use of medication 2005: 6.4%, 2004: 6.2% AOK: 11.4%; Barmer: 8.5%; DAK: 6.5%; BKK: 5.9%; TKK: 5.8%; PKV: 4.6%; IKK: 3.9%; other provision: 5.2% Registry data Diabetes register of the former GDR; 1960–1989 Physician-diagnosed 1989: 4.1% All age groups, total population [20, 39] diabetes 1987: 4.0% 1960: 0.6% Diabetes register of the former GDR; 1960–1989 1988: 3.8% (women: 4.5%, men: 3.0%) All age groups, population of East Berlin [47] 1970: 2.5% Regional study Rural Health Study, sample from randomly selected north-east 2004/2008, Self-report on diabetes 2004/2008: women: 12.4%, men: German rural communities; 1994, 1973 diagnosis 12.8% 18+ years, N=1.246 (2004/08), N=2.155 (1994), N=3.603 (1973) (2004/08) 1994: women: 10.9%, men: 6.8% [48] 1973: women: 3.5%, men: 3.1% DIAB-CORE Consortium including GNHIES98 (1997–99) and 1997–2006 Self-report on diabetes Type-2: 8.6% the regional studies CARLA (2002–06), DO-GS (2003/04), HNR (31/12/2007) diagnosis or anti-diabetic GNHIES98: 8.2%; CARLA: 12.0%; DO- (2000–03), KORA S4 (1999–2001), SHIP (1997–2001); medication and age at GS: 9.3%; HNR: 7.2%; KORA: 5.8%; SHIP: 45–74 years, N=15.071 [32] diagnosis >30 Years 10.9% KORA F4 (follow-up), cohort study in the Augsburg region; 2006–2008 Self-report on physician- Overall: 2.2% (women: 2.3%, men: 35–59 years, N=1.653 [33] (31/12/2007) diagnosed diabetes or 2.2%) KORA S4, survey in the Augsburg region; 1999–2001 anti-diabetic medication Overall: 8.7% (Women: 8.0%, Men: 55–74 years, N=1.353 [34] (31/12/2000) (verified) 9.3%)

6 | Bundesgesundheitsblatt - Gesundheitsforschung - Gesundheitsschutz 5/6 · 2013 Tab. 6 Overview of studies in Germany with data on the prevalence of known diabetes in adults (Continued) Study population Period (Stan- Definition of diabetes Prevalence of diabetes dardisation) MONICA/KORA, surveys in the Augsburg region; 1999–2001, Self-report on physician- 25–64 years: 25–64 years (19984/85), 25–74 years (subsequent surveys), 1994/95, diagnosed diabetes or 1999–2001: women: 2.7%, men: 2.5% N≈2.000 (per survey) [35] 1989/90, anti-diabetic medication 1994/1995: women: 2.5%, men: 3.0% 1984/85 1989/1990: women: 2.3%, men: 3.6% (31/12/2000) 1984/85: women: 2.0%, men: 2.7% 25–74 years: 1999–2001: women: 3.5%, men: 3.7% 1994/1995: women: 3.7%, men: 4.1% 1989/90: women: 3.6%, men: 4.6% EPIC-Potsdam, baseline examination of a cohort study in the 1994–1998 Self-report on physician- Women: 2.6%, men: 3.7% Potsdam region; (2007) diagnosed diabetes or 35–59 years, N≈27.500 [11] diabetic therapy Sample from randomly selected towns and rural areas in 5 federal 1993–1996 Self-report on diabetes Overall: 6.3% states; diagnosis or diabetes 18–70 years, N=2.150 [36] therapy Study on early detection of diabetes in Munich; 1967/1968 Self-report on diabetes Overall: 2.0% All age groups, N≈790.000 [49] diagnosis Health insurance fund data AOK Hesse/KV Hesse Statutory Health Insurance Sample; 2000–2009 Physician-diagnosed 2009: 9.8% All age groups, N≈300.000 per annum [5] (31/12/2009) diabetes (in ≥3 of 4 2000: 7.5% AOK Hesse/KV Hesse Statutory Health Insurance Sample; 2000–2007 quarters) or prescription 2007: 8.9% All age groups, N≈300.000 per annum [50] (31/12 of the of anti-diabetic medica- 2000: 6.5% previous year) tion (≥2 per annum or 1 per annum plus diabetes AOK Hesse/KV Hesse Statutory Health Insurance Sample; 1998–2004 2004: 7.9% (women: 8.1%, men 7.6%) diagnosis or plus glucose- All age groups, N≈300.000 per annum [21, 22] (31/12 of the 2001: 6.9% (women: 7.4%, men: 6.5%) or HbA1c-measurement previous year) 1998: 5.9% in the same quarter) Sample of insured persons from six statutory health insurance 1999 Diabetes diagnosis or pre- Overall: 6.5% funds; scription of anti-diabetic All age groups, N=14.7 million [23] medication Sample of insured persons from the AOK-Dortmund; 1988 Physician-diagnosed Overall: 4.8% (women: 5.5%, men: All age groups, N=6.478 [24] (1988) diabetes (in ≥2 of 4 quar- 4.1%) ters) or prescription of anti-diabetic medication (≥4 per annum) or blood glucose measurement (in ≥3 of 4 quarters) Patient data GEMCAS, national sample of patients from general practices; 2005 Physician-diagnosed Overall: 11.8% 18+ years, N=35.869 (N=1.511 practices) [25] (2003) diabetes Type-1: 0.7%, Type-2: 11.1% DETECT study, national sample of patients from general practices; 2003 Physician data on dia- Overall: 15.2% 18+ years, N=55.518 (N=3.188 practices) [37] betes diagnosis on day Type-1: 0.5%, Type-2: 14.7% of study or anti-diabetic medication ESTHER, sample of patients from health check-ups in general 2000–2002 Physician-diagnosed Overall: 13.8% practices in the Saarland; diabetes or anti-diabetic 50–74 years, N=9.953 [51] medication HYDRA-study, national sample of patients from general practices/ 2001 Physician data on diabe- Overall:15.6% (women: 13.7%, men: internists in private practice; tes diagnosis on day of 18.5%) 16+ years, N=43.549 (N=1.912 practices) [26] study SESAM 2, sample of patients from general practices in Saxony; 1999/2000 Physician-diagnosed Overall: 9.2% 2–102 years, N=8.877 (N=270 practices) [38] (2000) diabetes Type-1: 0.3%, Type-2: 6.2%, Other types: 1.8%, types not further speci- fied: 1.0% aDeviation in the data in Thefeld et al. [30] from . Tab. 5 is a result of the non-consideration of anti-diabetic medication in the definition of diabetes

Bundesgesundheitsblatt - Gesundheitsforschung - Gesundheitsschutz 5/6 · 2013 | 7 Main topic

shows no increase in prevalence for the Factors relevant to the The prevalence of known diabetes— 25- to 69-year-old population between interpretation of the prevalence and also that of undiagnosed diabetes—is 1990–1992 and 1997–1999 [30]—neither of known diabetes furthermore dependent on the incidence for the new nor for the old federal states and mortality rates. The incidence rate is [40]. Furthermore, no trend can be seen The magnitude of known diabetes is de- closely linked with behaviour (such as di- in an extension of this comparison over pendent on the ratio of diagnosed-to- et, physical activity and associated body time that includes the data from the tele- undiagnosed cases. Isolated analyses to weight) and living conditions (such as phone health surveys of 2002–2005 [40]. estimate this ratio lead to different re- personal or regional economic factors, air A comparison of the DEGS1 data with sults owing to differing study popula- pollution), which play a role in the devel- those from GNHIES98 as described in tions and study periods and especially opment of diabetes [2]. This means that the results section, indicates, in agree- owing to varying diagnosis criteria. Re- the observed increase in [43], the ment with other current investigations, sults of the oral glucose tolerance test as main risk factor for diabetes, will proba- an increase in the prevalence of known part of the KORA Study in the Augsburg bly as a consequence lead to an increase diabetes within the last decade. The in- region indicate a prevalence of undiag- in the incidence of diabetes. In addition crease for the 18- to 79-year-old popula- nosed diabetes (defined as fasting glucose to general life expectancy, the mortality tion amounts to 38% on the basis of the ≥7.0 mmol/l or glucose following the oral rate of persons with diabetes is associated examination surveys performed between glucose tolerance test ≥11.1 mmol/l) that with treatment intensity or rather treat- 1997–1999 and 2008–2011 and 29% on is just as high as that of diagnosed dia- ment success. Thus, given a higher gen- the basis of the Bertelsmann Healthcare betes [33, 34]. On the contrary, prelimi- eral life expectancy and improved treat- Monitor survey between 2004 and 2008 nary evaluations from DEGS1, in which ment possibilities, a higher prevalence [27]. For adults aged 18 years and above, no oral glucose tolerance test was con- would also be expected. So far, popula- an increase of 43% can be observed based ducted but in which laboratory values ex- tion-based data on incidence and mor- on the telephone surveys between 2002– ist for fasting glucose (48% of the sample) tality—other than from the diabetes reg- 2003 and 2008–2010 and 49% based on or random glucose and for HbA1c, indi- ister of the former GDR—are only avail- the AOK Hesse/KV Hesse sample of in- cate a prevalence of undiagnosed diabe- able from the Augsburg region [44, 45]. sured persons between 2000 and 2009 tes (defined as HbA1c ≥6.5% or fasting- The monitoring data of the RKI will [5]. A greater increase is evident in men glucose ≥7.0 mmol/l or random-glucose also be able to contribute towards esti- than in women: In the examination sur- ≥11.1 mmol/l) [42] that is more than three mating nationwide diabetic incidence veys this amounts to 49% versus 30%, in times lower than that of known diabetes. and mortality and temporal trends there- the telephone surveys 54% versus 34% In earlier studies in which no oral glucose in. and in the sample of insured persons tolerance test was conducted, the preva- 57% versus 40%. In addition, data analy­ lence of undiagnosed diabetes based on Strengths and limitations sis of the examination surveys shows—in various criteria was also substantially accordance with the results of the AOK lower than the prevalence of diagnosed The strengths of this study lie in the pop- Hesse/KV Hesse sample of insured per- diabetes [25, 30, 36]. ulation representativeness of the preva- sons [5]—that approximately one third of The ratio of diagnosed-to-undiag- lence estimates for the German resident the increase observed can be attributed to nosed diabetes can shift over the course population aged 18–79 years, and in the demographic ageing within the observa- of time. Thus, improved screening meth- comparability of the prevalence figures tion period. ods, increased attention by doctors (for at two points in time more than 10 years Projections of the prevalence of known example through the introduction of dis- apart due to identical sampling and def- type-2 diabetes for the year 2030—which ease management programmes) and im- inition of known diabetes. Estimates of are calculated on the basis of the diabetes proved knowledge about symptoms on prevalence and extrapolations for the prevalence data of population-based re- the part of the patients all lead to the ear- population are made possible by weight- gional studies and of GNHIES98 (DIAB- lier discovery of persons with diabetes. ing the results with complex weight- CORE Consortium), diabetes incidence Consequently, the proportion of diag- ing factors, which take into account the and mortality data from the Augsburg re- nosed cases increases whilst the propor- complex design of the study, the non-re- gion (KORA S4/F4) and the Federal Sta- tion of undiagnosed cases falls. Thus, da- sponse, and re-participation of the GN- tistical Office population forecast—show ta on the trend in undiagnosed diabetes, HIES98 participants [15, 16]. The lim- an increase of approximately 1.5 million which are gathered over time using com- itations of this study are that the popu- persons for the 55- to 74-year-olds alone parable methods, are essential for the in- lation aged 80 years and above and cer- (+64%) compared to today [41]. A more terpretation of the trends observed in tain groups of people (especially care- pronounced increase of almost 1 million known diabetes. However, there are no home residents and severely ill persons) people was projected for men (+79%) such data yet for Germany [11]. Planned could not be included representatively in than for women at over 0.5 million peo- comparative analyses of DEGS1 and GN- the examination survey. In addition, no ple (+47%). HIES98 will be able to provide an indica- comprehensive assessment of disease de- tion of this for the first time. velopment is possible since the estima-

8 | Bundesgesundheitsblatt - Gesundheitsforschung - Gesundheitsschutz 5/6 · 2013 tions here are limited to already diag- 16. Kamtsiuris P, Lange M, Hoffmann R et al (2013) References The first wave of the German Health Interview nosed diabetes. Furthermore, the defini- and Examination Survey for Adults (DEGS1). Sam- tion of known diabetes is mainly based on 1. American Diabetes Association (2012) Diagnosis pling design, response, sample weights, and rep- and classification of diabetes mellitus. Diabetes self-reports by participants, which could, resentativeness. Bundesgesundheitsblatt Ge- Care 35(Suppl 1):64–71 sundheitsforschung Gesundheitsschutz 56:620– however, be evaluated as a relatively valid 2. Chen L, Magliano DJ, Zimmet PZ (2012) The 630 source for the purpose of ascertaining di- worldwide epidemiology of type 2 diabetes mel- 17. Robert Koch-Institut (2009) DEGS: Studie zur Ge- litus—present and future perspectives. Nat Rev agnosed diabetes [46]. sundheit Erwachsener in Deutschland. Projektbe- Endocrinol 8:228–236 schreibung. Beiträge zur Gesundheitsberichter- 3. Giani G, Janka HU, Hauner H et al (2004) Epide- stattung des Bundes. RKI, Berlin miologie und Verlauf des Diabetes mellitus in 18. Gößwald A, Lange M, Dölle R, Hölling H (2013) Conclusion and outlook Deutschland. In: Scherbaum WA, Kiess W (eds) The first wave of the German Health Interview Evidenzbasierte Leitlinie DDG and Examination Survey for Adults (DEGS1). Par- 4. Schunk M, Reitmeir P, Schipf S et al (2012) Health- ticipant recruitment, fieldwork, and quality man- The results of DEGS1 and of other re- related quality of life in subjects with and with- agement. Bundesgesundheitsblatt Gesundheits- out Type 2 diabetes: pooled analysis of five pop- cent nationwide surveys indicate a fig- forschung Gesundheitsschutz 56:611–619 ulation-based surveys in Germany. Diabet Med 19. Lampert T, Kroll L, Müters S, Stolzenberg H (2013) ure of at least 4.6 million 18- to 79-year- 29:646–653 Measurement of Socioeconomic Status in the old adults in Germany having been di- 5. Koster I, Schubert I, Huppertz E (2012) Follow up German Health Interview and Examination Sur- of the CoDiM-Study: cost of diabetes mellitus agnosed with diabetes at some point. vey for Adults (DEGS1). Bundesgesundheits­ 2000–2009. Dtsch Med Wochenschr 137:1013– blatt Gesundheitsforschung Gesundheitsschutz Particularly affected are persons aged 1016 56:631–636 50 years and older, as well as peo- 6. Danaei G, Finucane MM, Lu Y et al (2011) Nation- 20. Michaelis D, Jutzi E (1991) Epidemiologie des Di- al, regional, and global trends in fasting plasma ple with a low socioeconomic status. abetes mellitus in der Bevölkerung der ehema- glucose and diabetes prevalence since 1980: sys- ligen DDR: Alters- und geschlechtsspezifische Within the last decade, the number tematic analysis of health examination surveys Inzidenz- und Prävalenztrends im Zeitraum of adults with known diabetes has in- and epidemiological studies with 370 country- 1960–1987. Z Klin Med 46:59–64 years and 2.7 million participants. Lancet 378:31– creased by 38%, of which approximate- 21. Hauner H, Köster I, Schubert I (2007) Prevalence 40 of diabetes mellitus and quality of care in Hesse, ly one third can be attributed to the de- 7. International Diabetes Federation (2011) IDF Dia- Germany, 1998–2004. Dtsch Arztebl 104:A2799– mographic ageing of the population. betes Atlas, 5th edn. International Diabetes Fed- A2805 eration, Brussels, Belgium Planned analy­ses on the temporal trend 22. Hauner H, Koster I, Ferber L von (2003) Preva- 8. Heidemann C, Du Y, Scheidt-Nave C (2011) Dia- lence of diabetes mellitus in Germany 1998– in undiag­nosed diabetes and on the in- betes mellitus in Germany. In: GBE kompakt. Ro­ 2001. Secondary data analysis of a health insur- cidence and mortality rates will con- bert-Koch-Institut, Berlin. http://www.rki.de/EN/ ance sample of the AOK in Hesse/KV in Hesse. Content/Health_Monitoring/Health_Reporting/ tribute to the interpretation of these re- Dtsch Med Wochenschr 128:2632–2637 Kompakt/Kompakt_node.html 23. Stock SA, Redaelli M, Wendland G et al (2006) Di- sults. 9. Robert Koch-Institut (RKI) (2010) Daten und abetes—prevalence and cost of illness in Ger- Fakten: Ergebnisse der Studie “Gesundheit in many: a study evaluating data from the statu- Deutschland aktuell 2009”. RKI, Berlin tory health insurance in Germany. Diabet Med Corresponding address 10. Robert Koch-Institut (RKI) (2012) Daten und 23:299–305 Fakten: Ergebnisse der Studie “Gesundheit in 24. Hauner H, Ferber L von, Koster I (1992) Estima- Deutschland aktuell 2010”. RKI, Berlin Dr. C. Heidemann tion of the incidence of diabetes in the Federal 11. Schulze MB, Rathmann W, Giani G, Joost HG Republic of Germany based on insurance data. A Department of Epidemiology and Health (2010) Diabetesprävalenz – Verlässliche Daten ­ Monitoring, Robert Koch Institute secondary data analysis of a representative ran- stehen noch aus. Dtsch Arztebl 107:A1694– dom sample of locally insured persons in the city General-Pape-Str. 62–66, 12101 Berlin A1696 of Dortmund. Dtsch Med Wochenschr 117:645– Germany 12. Kurth BM, Lange C, Kamtsiuris P, Hölling H (2009) 650 [email protected] Health monitoring at the Robert Koch Institute. 25. Hauner H, Hanisch J, Bramlage P et al (2008) Status and perspectives. Bundesgesundheits­ Prevalence of undiagnosed Type-2-diabetes mel- blatt Gesundheitsforschung Gesundheitsschutz litus and impaired fasting glucose in German pri- Acknowledgements. We would like to thank 52:557–570 mary care: data from the German Metabolic and R. Paprott for assisting with the literature research. The 13. Kurth BM (2012) Das RKI-Gesundheitsmonitor- Cardiovascular Risk Project (GEMCAS). Exp Clin study was financed with resources from the ­Robert ing – was es enthält und wie es genutzt werden Endocrinol Diabetes 116:18–25 Koch Institute and the Federal Ministry of Health. This kann. Public Health Forum 20(76):4.e1–4.e3 26. Lehnert H, Wittchen HU, Pittrow D et al (2005) work was supported by funding from the Kompetenz­ 14. Gößwald A, Lange M, Kamtsiuris P, Kurth BM Prevalence and pharmacotherapy of diabetes netz Diabetes mellitus (Competence Network Diabe- (2012) DEGS: German Health Interview and Ex- mellitus in primary care. Dtsch Med Wochenschr tes mellitus) funded by the Federal Ministry of Educa- amination Survey for Adults. A nationwide cross- 130:323–328 tion and Research (FKZ: 01GI1110F). sectional and longitudinal study within the 27. Hoffmann F, Icks A (2011) Diabetes prevalence framework of health monitoring conducted by based on health insurance claims: large differenc- Conflict of interest. On behalf of all authors, the the Robert Koch Institute. Bundesgesundheits- es between companies. Diabet Med 28:919–923 corresponding author states the following: With- blatt Gesundheitsforschung Gesundheitsschutz 28. Hoffmann F, Icks A (2012) Diabetes ‘Epidem- in the scope of lectures and consultancy work, Dr. 55:775–780 ic’ in Germany? A critical look at health insur- W. Rathmann has received fees from the following 15. Scheidt-Nave C, Kamtsiuris P, Gößwald A et al ance data sources. Exp Clin Endocrinol Diabetes companies: BMS, Eli Lilly, NovoNordisk, IMS HEALTH (2012) German health interview and examination 120(7):410–415 and Sanofi-Aventis. Dr. C. Heidemann, Dr. Y. Du, Dr. I. survey for adults (DEGS)—design, objectives and 29. Hoffmann F, Icks A (2012) Structural differences Schubert, and Dr. C. Scheidt-Nave declare that there is implementation of the first data collection wave. between health insurance funds and their impact no conflict of interest. BMC Public Health 12:730 on health services research: results from the Ber- telsmann Health-Care Monitor. Gesundheitswe- sen 74:291–297 30. Thefeld W (1999) Prevalence of diabetes mellitus in the adult German population. Gesundheitswe- sen 61(Spec No):85–89

Bundesgesundheitsblatt - Gesundheitsforschung - Gesundheitsschutz 5/6 · 2013 | 9 Main topic

31. Helmert U, Janka HU, Strube H (1994) Epidemio- 46. Midthjell K, Holmen J, Bjorndal A, Lund-Larsen G logic findings for diabetes mellitus prevalence in (1992) Is questionnaire information valid in the the Federal Republic of Germany between 1984 study of a chronic disease such as diabetes? The and 1991. Diab Stoffw 3:271–277 Nord-Trondelag diabetes study. J Epidemiol Com- 32. Schipf S, Werner A, Tamayo T et al (2012) Region- munity Health 46:537–542 al differences in the prevalence of known type 47. Ratzmann KP (1991) An analysis og age-specif- 2 diabetes mellitus in 45–74 years old individu- ic and sex-specific diabetes prevalence and its als: results from six population-based studies in mode of treatment: The Berlin Study. Akt Endokr Germany (DIAB-CORE Consortium). Diabet Med Stoffw 12:220–223 29(7):e88–e95 48. Elkeles T, Beck D, Roding D et al (2012) Health 33. Meisinger C, Strassburger K, Heier M et al (2010) and lifestyle in rural northeast Germany: the find- Prevalence of undiagnosed diabetes and im- ings of a Rural Health Study from 1973, 1994, and paired glucose regulation in 35–59-year-old indi- 2008. Dtsch Arztebl Int 109:285–292 viduals in Southern Germany: the KORA F4 Study. 49. Mehnert H, Sewering H, Reichstein W, Vogt H Diabet Med 27:360–362 (1968) Early detection of diabetics in Munich 34. Rathmann W, Haastert B, Icks A et al (2003) High 1967–68. Dtsch Med Wochenschr 93:2044–2050 prevalence of undiagnosed diabetes mellitus in 50. Köster I, Huppertz E, Hauner H, Schubert I (2011) Southern Germany: target populations for effi- Direct Costs of Diabetes Mellitus in Germany— cient screening. The KORA survey 2000. Diabeto- CoDiM 2000–2007. Exp Clin Endocrinol Diabetes logia 46:182–189 119:377–385 35. Meisinger C, Heier M, Doering A et al (2004) Prev- 51. Maatouk I, Wild B, Wesche D et al (2012) Tempo- alence of known diabetes and antidiabetic thera- ral predictors of health-related quality of life in py between 1984/1985 and 1999/2001 in south- elderly people with diabetes: results of a german ern Germany. Diabetes Care 27:2985–2987 cohort study. PloS One 7:e31088 36. Palitzsch K-D, Nusser J, Arndt H et al (1999) The prevalence of diabetes mellitus is significantly underestimated in Germany, according to a na- tionwide epidemiological study based on HbA1c analysis. Diab Stoffw 8:189–200 37. Pittrow D, Stalla GK, Zeiher AM et al (2006) Prev- alence, drug treatment and metabolic control of diabetes mellitus in primary care. Med Klin (Mu- nich) 101:635–644 38. Frese T, Sandholzer H, Voigt S, Voigt R (2008) Ep- idemiology of diabetes mellitus in German gen- eral practitioners’ consultation—results of the SESAM 2-study. Exp Clin Endocrinol Diabetes 116:326–328 39. Michaelis D, Jutzi E, Vogt L (1993) Epidemiology of insulin-treated diabetes mellitus in the East- German population: differences in long-term trends between incidence and prevalence rates. Diabete Metab 19:110–115 40. Heidemann C, Kroll L, Icks A et al (2009) Preva- lence of known diabetes in German adults aged 25–69 years: results from national health surveys over 15 years. Diabet Med 26:655–658 41. Brinks R, Tamayo T, Kowall B, Rathmann W (2012) Prevalence of type 2 diabetes in Germany in 2040: estimates from an epidemiological model. Eur J Epidemiol 27:791–797 42. Heidemann C, Du Y, Scheidt-Nave C (2012) Diabe- tes mellitus: How many adults in Germany have diabetes? Bundesgesundheitsblatt Gesundheits- forschung Gesundheitsschutz 55:983–984 43. Mensink GB, Schienkiewitz A, Haftenberger M et al (2013) Overweigt and obesity in Germany. Re- sults of the German Health Interview and Exam- ination Survey for Adults (DEGS1). Bundesge- sundheitsblatt Gesundheitsforschung Gesund- heitsschutz 56:786–794 44. Rathmann W, Strassburger K, Heier M et al (2009) Incidence of Type 2 diabetes in the elderly Ger- man population and the effect of clinical and life- style risk factors: KORA S4/F4 cohort study. Dia- bet Med 26:1212–1219 45. Kowall B, Rathmann W, Heier M et al (2011) Cat- egories of glucose tolerance and continuous gly- cemic measures and mortality. Eur J Epidemiol 26:637–645

10 | Bundesgesundheitsblatt - Gesundheitsforschung - Gesundheitsschutz 5/6 · 2013