[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 54.205.176.107. Please contact the publisher to request reinstatement.
[Skip to Content Landing]
Download PDF
Table. Performances of the Tested Scores
Table. Performances of the Tested Scores
1.
Jönsson B.CODE-2 Advisory Board.  Revealing the cost of type II diabetes in Europe.  Diabetologia. 2002;45(7):S5-S12PubMedArticle
2.
Wild S, Roglic G, Green A, Sicree R, King H. Global prevalence of diabetes: estimates for the year 2000 and projections for 2030.  Diabetes Care. 2004;27(5):1047-1053PubMedArticle
3.
Kahn HS, Cheng YJ, Thompson TJ, Imperatore G, Gregg EW. Two risk-scoring systems for predicting incident diabetes mellitus in US adults age 45 to 64 years.  Ann Intern Med. 2009;150(11):741-751PubMed
4.
Balkau B, Lange C, Fezeu L,  et al.  Predicting diabetes: clinical, biological, and genetic approaches: data from the Epidemiological Study on the Insulin Resistance Syndrome (DESIR).  Diabetes Care. 2008;31(10):2056-2061PubMedArticle
5.
Griffin SJ, Little PS, Hales CN, Kinmonth AL, Wareham NJ. Diabetes risk score: towards earlier detection of type 2 diabetes in general practice.  Diabetes Metab Res Rev. 2000;16(3):164-171PubMedArticle
6.
Lindström J, Tuomilehto J. The diabetes risk score: a practical tool to predict type 2 diabetes risk.  Diabetes Care. 2003;26(3):725-731PubMedArticle
7.
Swiss Diabetes Association.  Diabète de type 2—Quel est votre risque? http://www.diabetesgesellschaft.ch/fr/informations/test-diabete/. Accessed January 10, 2011
8.
Wilson PW, Meigs JB, Sullivan L, Fox CS, Nathan DM, D’Agostino RB Sr. Prediction of incident diabetes mellitus in middle-aged adults: the Framingham Offspring Study.  Arch Intern Med. 2007;167(10):1068-1074PubMedArticle
9.
Firmann M, Mayor V, Vidal PM,  et al.  The CoLaus study: a population-based study to investigate the epidemiology and genetic determinants of cardiovascular risk factors and metabolic syndrome.  BMC Cardiovasc Disord. 2008;8:6PubMedArticle
Research Letters
Jan 23, 2012

Validation of 7 Type 2 Diabetes Mellitus Risk Scores in a Population-Based Cohort: CoLaus Study

Author Affiliations

Author Affiliations: Faculty of Biology and Medicine (Mr Schmid and Drs Vollenweider, Bastardot, Waeber, and Marques-Vidal), Institute of Social and Preventive Medicine (IUMSP) (Dr Marques-Vidal), and Division of Internal Medicine, Department of Medicine (Drs Vollenweider, Bastardot, and Waeber), Centre Hospitalier Universitaire Vaudois (CHUV), Lausanne, Switzerland. Mr Schmid is a medical student at IUMSP, CHUV.

Arch Intern Med. 2012;172(2):188-189. doi:10.1001/archinte.172.2.188

One of the challenges for public health in the coming years is the expected increase of type 2 diabetes mellitus (T2DM) prevalence and its resulting health burden and costs.1,2 For the physician, although recommendations regarding who to screen for T2DM are available, the application of a validated risk score would enable a better targeting of high-risk subjects and thus an improvement of preventive measures. Indeed, numerous risk scores for T2DM have been developed, but few studies have compared them in populations different from those they have been derived from. It is also unclear whether all risk scores have the same prognostic validity.

The aim of this study was to assess the validity of various T2DM risk scores in predicting the incidence of T2DM in a Swiss population-based cohort.

Methods

Seven T2DM risk scores were selected in the present study. Four were based on clinical data: the 10-year risk score from Kahn et al3 (Kahn clinical); the 9-year risk score from Balkau et al4; the prevalent undiagnosed diabetes risk score from Griffin et al5; the Finnish Type 2 Diabetes Risk Score (FINDRISC), which has been developed in 2 cohorts followed for 5 and 10 years6; and finally the risk score from the Swiss Diabetes Association, available online,7 which is actually adapted from FINDRISC. The 2 remaining risk scores were based on the association of clinical and biological data: the 10-year risk score from Kahn et al3 (Kahn clinical + biologic) and the 8-year risk score from Wilson et al.8 We used the thresholds provided by the authors, and each score had its area under the receiver operating characteristic curve (AROC), sensitivity, specificity, and negative and positive predictive values assessed. We tested these scores in 3060 nondiabetic participants from Lausanne, Switzerland (44.6% men; mean [SD] age, 52.6 [10.6] years), followed up for 5 years (study period, 2003-2011).9 Incident diabetes was defined as fasting plasma glucose level greater than or equal to 126.13 mg/dL (to convert to millimoles per liter, multiply by 0.0555) and/or presence of oral hypoglycemic or insulin treatment.

Results

A total of 169 patients (5.5%) developed T2DM during follow-up. Compared with participants who did not develop T2DM, they were more frequently male (69.8% vs 43.1%); were older (mean [SD] age, 57.1 [9.4] vs 52.3 [10.6] years); had a higher frequency of family history of T2DM (31.4% vs 19.3%) (all P < .001); and had a higher resting heart rate (69 [10] vs 67 [9] beats/min [P < .05]). They practiced less leisure-time physical activity (45.6% vs 60.3%); had higher body mass index (29.0 [3.9] vs 25.1 [4.0] [calculated as weight in kilograms divided by height in meters squared]), waist circumference (100.3 [10.9] vs 86.8 [12.2] cm), and fasting plasma glucose (110.45 [9.37] vs 95.32 [9.37] mg/dL), triglyceride (189.38 [184.07] vs 111.50 [79.65] mg/dL [to convert to millimoles per liter, multiply by 0.0113]), and uric acid (6.03 [1.32] vs 5.11 [1.35] mg/dL [to convert to micromoles per liter, multiply by 59.485]) levels; and had lower high-density lipoprotein cholesterol levels (53.28 [13.51] vs 64.09 [16.60]mg/dL [to convert to millimoles per liter, multiply by 0.0259]) (all P < .001). The performance of the 7 T2DM risk scores is given in the Table. Most risk scores had a high AROC, specificity, and negative predictive value, while their sensitivity and positive predictive values were low.

Comment

Most variables included in the risk scores were significantly different between participants who developed T2DM and those who did not, which confirms their prognostic role. The best results were obtained by the Kahn clinical + biologic risk score. However, a risk score based on simple clinical data (FINDRISC) also had a high AROC, which could be more convenient regarding health costs and acceptability by patients. Indeed, using data from our hospital, applying the Kahn clinical + biologic risk score would lead to an extra cost of US$ 12.02 per screened patient relative to the FINDRISC score.

Our study has several limitations. Follow-up time was limited to 5 years; still, our findings are in agreement with the performances reported in the original studies, suggesting that our results should also be reliable after a 10-year follow-up. Some factors such as fruit consumption and second-degree familial history could not be assessed in this study owing to lack of information; although we corrected for such missing data, it is possible that the performance of the corresponding risk scores might have been reduced. Still, one of these risk scores (FINDRISC) ranked second best in our study, suggesting that the reduction in its predictive power may not be significant. In this study, physical activity was defined as at least 2 h/wk of leisure-time physical activity, but it was defined as 4 h/wk or 30-min/d in the original publications.6,7 Finally, this study was limited to white participants and whether the results also apply to other ethnicities is unknown.

In conclusion, this is the first study, to our knowledge, to compare the prognostic validity of several risk scores for T2DM. The Kahn clinical + biologic risk score has the highest AROC, but the clinical FINDRISC score may be more practical and less expensive for screening. Further research is needed to assess the real impact of these scores in preventing T2DM.

Back to top
Article Information

Correspondence: Dr Marques-Vidal, Institut Universitaire de Médecine Sociale et Préventive, 17, rue du Bugnon, 1005 Lausanne, Switzerland (Pedro-Manuel.Marques-Vidal@chuv.ch).

Author Contributions:Study concept and design: Schmid, Vollenweider, Waeber, and Marques-Vidal. Acquisition of data: Vollenweider, Bastardot, and Waeber. Analysis and interpretation of data: Schmid and Marques-Vidal. Drafting of the manuscript: Schmid. Critical revision of the manuscript for important intellectual content: Vollenweider, Bastardot, Waeber, and Marques-Vidal. Statistical analysis: Schmid. Obtained funding: Vollenweider and Waeber. Administrative, technical, and material support: Bastardot. Study supervision: Marques-Vidal.

Financial Disclosure: Drs Vollenweider and Waeber received an unrestricted grant from GlaxoSmithKline to build the CoLaus study.

Funding/Support: The CoLaus study was supported by research grants from the Swiss National Science Foundation (grant 33CSCO-122661) from GlaxoSmithKline and the Faculty of Biology and Medicine of Lausanne, Switzerland.

Additional Contributions: We thank the participants in the Lausanne CoLaus study and the investigators who have contributed to the recruitment, in particular research nurses Yolande Barreau, Anne-Lise Bastian, Binasa Ramic, Martine Moranville, Martine Baumer, Marcy Sagette, Jeanne Ecoffey, Sylvie Mermoud, Nicole Bonvin, Laure Bovy, Nathalie Maurer, Vanessa Jaquet, and Nattawan Laverrière for data collection. No compensation was received for all persons named in the acknowledgment.

References
1.
Jönsson B.CODE-2 Advisory Board.  Revealing the cost of type II diabetes in Europe.  Diabetologia. 2002;45(7):S5-S12PubMedArticle
2.
Wild S, Roglic G, Green A, Sicree R, King H. Global prevalence of diabetes: estimates for the year 2000 and projections for 2030.  Diabetes Care. 2004;27(5):1047-1053PubMedArticle
3.
Kahn HS, Cheng YJ, Thompson TJ, Imperatore G, Gregg EW. Two risk-scoring systems for predicting incident diabetes mellitus in US adults age 45 to 64 years.  Ann Intern Med. 2009;150(11):741-751PubMed
4.
Balkau B, Lange C, Fezeu L,  et al.  Predicting diabetes: clinical, biological, and genetic approaches: data from the Epidemiological Study on the Insulin Resistance Syndrome (DESIR).  Diabetes Care. 2008;31(10):2056-2061PubMedArticle
5.
Griffin SJ, Little PS, Hales CN, Kinmonth AL, Wareham NJ. Diabetes risk score: towards earlier detection of type 2 diabetes in general practice.  Diabetes Metab Res Rev. 2000;16(3):164-171PubMedArticle
6.
Lindström J, Tuomilehto J. The diabetes risk score: a practical tool to predict type 2 diabetes risk.  Diabetes Care. 2003;26(3):725-731PubMedArticle
7.
Swiss Diabetes Association.  Diabète de type 2—Quel est votre risque? http://www.diabetesgesellschaft.ch/fr/informations/test-diabete/. Accessed January 10, 2011
8.
Wilson PW, Meigs JB, Sullivan L, Fox CS, Nathan DM, D’Agostino RB Sr. Prediction of incident diabetes mellitus in middle-aged adults: the Framingham Offspring Study.  Arch Intern Med. 2007;167(10):1068-1074PubMedArticle
9.
Firmann M, Mayor V, Vidal PM,  et al.  The CoLaus study: a population-based study to investigate the epidemiology and genetic determinants of cardiovascular risk factors and metabolic syndrome.  BMC Cardiovasc Disord. 2008;8:6PubMedArticle
×