Update: The ADVANCE risk engine for cardiovascular disease prediction in the context of current strategies for cardiovascular risk evaluation in people with diabetes

Andre Pascal KENGNE, MD, PhD
South African Medical Research Council and University of Cape Town
The George Institute for Global Health, Sydney – AUSTRALIA

The ADVANCE risk engine for cardiovascular disease prediction in the context of current strategies for cardiovascular risk evaluation in people with diabetes

by A. P. Kengne, South Africa and Australia

Until recently, cardiovascular disease (CVD) risk stratification in people with diabetes has been controversial. This review, based on evidence from the ADVANCE study (Action in Diabetes and Vascular disease: PreterAx and DiamicroN MR Controlled Evaluation) and relevant publications identified via PubMed, examines existing approaches to CVD risk evaluation in diabetics, and discusses the use of absolute CVD risk tools as an appropriate basis for CVD prevention in such people. Evidence shows that diabetes is not a CVD risk equivalent in all circumstances. In diabetics, CVD risk follows a gradient. Reliably capturing this gradient depends upon assessing an adequate combination of risk factors. Many global CVD risk tools applicable to diabetics have been developed. Those derived from older cohorts are less accurate in contemporary populations and many newer tools have not been tested. The ADVANCE risk engine, recently developed from the large multinational ADVANCE study, performed acceptably in the ADVANCE population, and largely outperformed the popular Framinghamrisk equation when tested on the multinational DIABHYCAR (non-insulin–dependent DIABetes, HYpertension, microalbuminuria or proteinuria, CARdiovascular events, and ramipril) cohort of type 2 diabetics. In conclusion, the high-risk status conferred by diabetes does not preclude estimation of absolute CVD risk using tools such as the ADVANCE risk engine and its use in initiating and intensifying preventative measures. Adopting such an accurate and validated tool will likely improve therapeutic choices and outcomes in diabetes care.

Medicographia. 2013;35:102-109 (see French abstract on page 109)

Cardiovascular disease (CVD) risk estimation is motivated by the need to identify individuals whose outcomes can be modified by further investigation, initiation, or intensification of risk-modifying therapies.1,2 CVD risk estimates are also used to educate patients about their future chances of experiencing a cardiovascular event, encouraging adoption of healthy lifestyles and adherence to prescribed disease-modifying therapies in order to mitigate risk. Physicians engaged in routine care of individuals with diabetes are likely interested in quantifying patient risk of experiencing any major CVD outcome over a reasonable time horizon, using an accurate and validated global CVD risk evaluation tool.3-5

Until very recently, strategies for CVD prevention in people with diabetes were guided by the principle that future chances of experiencing CVD in diabetics were similar in magnitude to those in nondiabetic individuals with existing CVD. Based on this principle, diabetics were eligible for risk-reducing medications, such as statins, without accounting for absolute CVD risk levels. This principle, however, was inspired by evidence from earlier cohort studies,6-8 which may no longer reflect the modern era of diabetes care. Indeed, subsequent and more recent studies have shown variable results, with indications that the presence of diabetes mellitus may not be a “CVD risk equivalent” in all circumstances.9 These new findings have refocused interest on the need for a multivariable approach to risk stratification for CVD prevention in diabetics. Such an approach is especially relevant in the current era, characterized by the gradual shift in diabetes mellitus management from a glucocentric focus to an intensive multifactorial strategy targeting reduction in the risk of both macrovascular and microvascular complications of diabetes.10,11

This review begins with a critical examination of the concept of diabetes mellitus as a CVD risk equivalent from various perspectives, and subsequently discusses the rationale and strategies for global CVD risk estimation in diabetics, emphasizing the specificities and limitations of such strategies. The discussion largely draws from new knowledge gained from cardiovascular disease risk modeling in the ADVANCE study (Action in Diabetes and Vascular disease: PreterAx and DiamicroN MR Controlled Evaluation),3,12 and builds on the most up-to-date, relevant, and key literature regarding the concepts underlying use of prognostic models for global cardiovascular risk assessment in people with diabetes.

The concept of diabetes as a CVD risk equivalent

The concept of CVD risk equivalence was inspired by a study by Haffner et al.6 That study showed a nonsignificant difference in the risk of coronary heart disease (CHD) death in diabetics with no prior myocardial infarction (MI) compared with their nondiabetic counterparts with a history of MI. Accordingly, some guidelines have advocated widespread use of statins and aspirin in diabetic patients following the standards applicable to nondiabetic patients with a history of CVD.13-15 The study by Haffner et al,6 however, was based on data collected between 1982 and 1990, well before the publication of landmark studies that have set the basis for contemporary diabetes care.16-20 Therefore, the study has many characteristics which may not reflect the profile of today’s diabetics. All study participants were white and were receiving glucose control medications; thus, the study population likely included more individuals with an advanced stage of the disease. This differs from the typical contemporary community-based cohort comprised of diabetics with various degrees of disease severity.3 Other limitations of the Haffner study include likelihood of misclassification of CHD events during follow-up and low statistical power. Subsequent studies, based on larger cohorts, did not necessarily corroborate the findings of Haffner and collaborators. Two independent meta-analyses of available relevant studies do not support the concept of CHD equivalence.9,21 Furthermore, there is wide variation in the rate of CHD in diabetes, depending on the population and coincident CVD risk factors.22

Recent data on the potential harms of preventative therapies (statins or aspirin) do not encourage “blanket” use of such therapies irrespective of a patient’s absolute CVD risk. Indeed, two meta-analyses of relevant statin trials23,24 and a subsequent individual study25 have shown that statin use increases the risk of hyperglycemia. Therefore, statins may adversely affect the outcomes of glucose-lowering strategies if used too liberally, particularly in low-risk individuals who may not necessarily derive cardiovascular benefits. The potential nonbeneficial effect of the routine use of aspirin for primary prevention was illustrated in a recent meta-analysis, showing that in people without prior CVD (including those with diabetes), aspirin does not reduce cancer mortality as previously thought; it also did not reduce cardiovascular death, and it induced clinically important bleeding events.26 This underscores the importance of considering an individual’s absolute risk in clinical decision-making regarding aspirin in people with diabetes.

An overview of global risk assessment

Global cardiovascular risk assessment is based on the combination of predictive information from several risk factors, using mathematical equations (models).2,27 In those models, the coefficient of each included risk factor indicates its relative contribution to the overall CVD risk.2,27 Once developed, a risk model normally requires validation both on the derivation sample (internal validation) and on independent populations (external validation). Validation consists of testing whether the model correctly estimates the risk of future events in one or several populations.2,27

The performance of absolute cardiovascular risk models are commonly assessed in terms of discrimination, calibration, and, more recently, reclassification.2,27 Discrimination is the ability of the model to correctly classify individuals who go on to develop a cardiovascular event and those who remain event-free.2,27 For example, if there are two individuals with diabetes, with one developing a cardiovascular event after a certain period of follow-up and the other remaining CVD free, a model with a high discriminative ability will systematically assign a higher risk to the first subject compared with the second. Discrimination is commonly characterized by the area under the receiver operating characteristic curve (AUC) or the C statistic. The C statistic ranges from 0.5 (lack of discrimination) to 1.0 (perfect discrimination).1,2,27 In general, a C statistic of 0.7 or greater is considered acceptable.

Calibration refers to the agreement between predicted risk and observed risk, and is assessed by comparing risk estimates from the model with actual event rates in the test population. 1,2,27 For instance, a 5-year estimated risk of cardiovascular disease of 20% for a patient means that, in a given group of patients with similar characteristics, 20% will experience a cardiovascular event within a 5-year period of follow up. The most commonly reported measure of calibration is the Hosmer-Lemeshow statistic. Estimates of calibration are sensitive to differing baseline levels of risk. For instance, if a given risk model is derived in a high-risk population, but tested in a low-risk population, the predicted risk estimates might be unreliably high. Recalibration of the risk model by adjusting the baseline risk estimates to fit the target population may help correct the overestimation or underestimation of risk.1,27

Global CVD risk estimation in diabetes: contemporary approaches

Given that diabetes is not a CVD risk equivalent in all circumstances, treatment decisions in diabetics should be based not on reduction in relative risk, but on reduction of absolute risk. Therefore, in patients with diabetes, estimation of global absolute CVD risk is of utmost importance. Three main approaches have been used to estimate global cardiovascular risk in people with diabetes.28

The first approach, based on the CVD risk equivalent concept, consists of classifying all individuals with diabetes as having a 10-year absolute CVD risk of at least 20%. However, given the aforementioned evidence, such an approach appears to be counterintuitive as CVD risk is not uniformly distributed among people with diabetes. This is corroborated by many studies,22,29 including a recent one showing multivariable risk prediction to be significantly better than classification of diabetes as a cardiovascular risk equivalent (10-year CVD risk of at least 20%).

The second approach consists of building unifying models for global CVD risk assessment for people with or without diabetes. The rationale for these tools is that there is no interaction between the diabetes status and other cardiovascular risk factors. In other words, everything else being equal, a subject with diabetes will not always have a higher risk than a nondiabetic subject with the same level of other risk factors (eg, blood pressure). This has been the basis for models such as the popular Framingham cardiovascular risk equations.28

The third approach, termed the interaction approach, consists of constructing separate models for people with and without diabetes. This approach assumes that risk factors affect cardiovascular disease risk in different ways in people with and without diabetes. A main limitation of this approach resides in the fact that models developed in one group cannot be used to inform risk stratification in the other group. This approach in people with diabetes was initially used by the UKPDS (United Kingdom Prospective Diabetes Study) investigators.30,31 This was based on the assumption that a unit increment in the duration of diagnosed diabetes contributes more to risk estimates than a unit increment in age.30 Therefore, to allow a more rational use of predictive information from age in people with diabetes, it has to be split into two components (ie, age at diabetes diagnosis and known duration of diabetes). While the assumption has not necessarily been tested and confirmed in other studies, this consideration has other useful applications.

Available studies largely suggest that classical cardiovascular risk factors, including smoking, blood pressure, lipid variables, and even some novel risk factors,28,32-35 affect the risk of CVD in similar ways in people with and without diabetes, with no evidence of interaction. Some risk factors or characteristics are likely to be more frequent in people with diabetes and may justify separate cardiovascular risk models for people with diabetes. These diabetes-specific characteristics include prescription of cardiovascular risk–reducing therapies, which may differ in people with and without diabetes. Additional specific factors, including glycated hemoglobin (HbA1c), urinary albumin excretion, and markers of microvascular complications of diabetes in general (especially retinopathy), have been demonstrated to be associated with CVD risk, and can contribute useful predictive information.36-41

Performance of popular CVD risk models and the ADVANCE study

At the time the ADVANCE study was conducted, cardiovascular risk prediction models in the general population were dominated by models developed from the Framingham Heart Study, of which many were applicable to diabetics.42 Diabetes specific models were also available, particularly those from the UKPDS study.42 However, the clinical utility and comparative performance of these models in contemporary populations with diabetes in diverse settings had not been established. Therefore, one major achievement was the implementation of extensive validation studies for popular existing cardiovascular risk models, using the unique features of the ADVANCE cohort.3 In the cohort of ADVANCE participants who had no known history of CVD at their enrolment in the trial, the 4-year risk of total cardiovascular events and particular components were largely overestimated by the Framingham-Anderson,43 Framingham-D’Agostino,44 and UKPDS equations.30,31 This overestimation was also observed in males and females, whites and non-whites, and the double placebo cohort (ie, those assigned to the placebo group in the blood pressure– lowering arm and the standard care group of the blood glucose– control arm).3 Discrimination of the Framingham and UKPDS equations in predicting CVD events in ADVANCE was poor for stroke, and modest-to-acceptable for CHD and total CVD. Recalibration substantially attenuated the magnitude of risk overestimation by the Framingham and UKPDS equations in ADVANCE. Discrimination, as expected, was unaffected, indicating the need for a new equation with improved discriminatory capability for people with diabetes, particularly those who are receiving many contemporary cardiovascular risk–reducing therapies.

Table I
Table I. Standardized (= ) coefficients (95% confidence intervals) and standard errors for predictors in the ADVANCE cardiovascular
disease prediction model.

Abbreviations: ADVANCE, Action in Diabetes and Vascular disease: PreterAx and DiamicroN MR Controlled Evaluation; CVD, cardiovascular disease; HbA1c, glycated hemoglobin; HDL, high-density lipoprotein; y, year.
After reference 12: Kengne et al. Eur J Cardiovasc Prev Rehabil. 2011;18(3):393-398. © 2011, European Society of Cardiology. http://online.sagepub.com.

Development of the ADVANCE model for cardiovascular prevention

In developing a new model for risk prediction, it is critical to account for the limitations of the existing ones in order to improve performance. In ADVANCE, the inclusion of participants from many countries provided the opportunity to take into account the substantial variation in the care of diabetes and CVD around the world, whereas other available models have been derived from homogeneous populations. The ADVANCE model targets total CVD and therefore captures the interrelation between components of CVD, such as CHD or stroke, unlike many existing models that focus specifically on these components. The complexity of the relationship between chronic hyperglycemia and cardiovascular risk is less fully addressed in existing models. Some improvement was achieved in the ADVANCE model through integration of risk factors to capture the exposure to chronic hyperglycemia both prior to and after the clinical diagnosis of diabetes. Statistical method is an important component of model development. Trusted statistical methods have been used to select potential risk factors and test their suitability for inclusion in the ADVANCE risk engine.12

Risk factors considered for inclusion in the ADVANCE model were age at clinical diagnosis of diabetes, duration of diagnosed diabetes, sex, blood pressure indices (systolic, diastolic, mean arterial, and pulse pressures), lipid variables (total, high-density lipoprotein [HDL], and non-HDL cholesterol; ratio of total/HDL cholesterol, and triglycerides), body mass index, waist circumference, waist-to-hip ratio, blood pressure– lowering medication (ie, treated hypertension), statin use, current smoking, retinopathy, atrial fibrillation (past or present), urinary albumin/creatinine ratio, serum creatinine, HbA1c, fasting blood glucose, and randomized treatments (blood pressure– lowering and glucose-control regimens). Ten of these candidate risk factors were included in the final ADVANCE model. To improve the applicability of the ADVANCE equation to other populations, age at diabetes diagnosis and known duration of diabetes were preferred to age at baseline. The β coefficients and accompanying standard error for risk factors in the ADVANCE cardiovascular risk model are shown in Table I.12

Performance of the ADVANCE risk model

The applicability of the ADVANCE model was tested on the same population used to develop the model (ie, internal validation) and on an independent external sample for which the DIABHYCAR (non-insulin–dependent DIABetes, HYpertension, microalbuminuria or proteinuria, CARdiovascular events, and ramipril participants) cohort45 was used (ie, external validation). In both internal and external validations, the discrimination of the ADVANCE model was acceptable. In comparison with existing total CVD models, the ADVANCE model largely did better at ranking the DIABHYCAR (event vs no event) than the Framingham-Anderson and Framingham- D’Agostino models. The calibration of the ADVANCE model was excellent in internal validation and good in external validation, with only a modest risk underestimation, likely explained by the difference in the levels of preventive therapies between ADVANCE and DIABHYCAR populations. Interestingly, the agreement between predictions by the ADVANCE models and the observed CVD events was consistent across different cut-offs for predicted risk of CVD. For comparison, the two Framingham equations overestimated the risk of CVD in the DIABHYCAR cohort by 65% (Anderson equation) and 99% (D’Agostino equation). Using a cut-off for 4-year predicted risk of >8%, which is approximately equivalent to a 10-year predicted risk of 20%and above, the ADVANCE model would reliably identify the 22% of the ADVANCE participants and 39% of the DIABHYCAR participants in whom 48% and 66% of CVD events, respectively, occurred during follow-up. Further intensifying treatment in such groups on top of any baseline therapy could achieve significant gain in terms of cardiovascular risk reduction.

Figure 1
Figure 1. Major cardiovascular disease points and 4-year predicted risk by the ADVANCE model equation.

As an illustration of the use of the risk scoring chart, a male subject, diagnosed with diabetes at the age of 50, who has a pulse pressure of 50 mm Hg and is currently treated for hypertension, and who since 3 years ago also has retinopathy, atrial fibrillation, and microalbuminuria, an HbA1c of 7%, and a non–HDL-C of 3.3 mmol/L
will receive a total score of 13 points: 0 for sex, 3 for age at diagnosis, 1 for known duration, 1 for pulse pressure, 1 for treated hypertension, 1 for retinopathy, 2 for
atrial fibrillation, 2 for microalbuminuria, and 1 each for HbA1c and non–HDL-C. A score of 13 points is equivalent to a 4-year estimated risk of 6.2%, which is similar
to the risk estimated for the same patient using the full equation.
Abbreviations: ADVANCE, Action in Diabetes and Vascular disease: PreterAx and DiamicroN MR Controlled Evaluation; CVD, cardiovascular disease; HbA1c, glycated
hemoglobin; HDL-C, high-density lipoprotein cholesterol.
After reference 12: Kengne et al. Eur J Cardiovasc Prev Rehabil. 2011;18(3):393-398. © 2011, European Society of Cardiology. http://online.sagepub.com.

Figure 2
Figure 2. Correlation
between risk
estimates from the
full ADVANCE cardiovascular
equation and those
from the handheld
(left panel) and online
(right panel)

ADVANCE, Action in
Diabetes and Vascular
disease: PreterAx and
DiamicroN MR Controlled

Dissemination of the ADVANCE model

To facilitate the uptake of the ADVANCE model in clinical practice, a handheld calculator and a risk scoring chart (Figure 1) have been developed.12 Other tools from this model, including an online calculator, are available on the model’s Web site, making it easily accessible and encouraging its adoption.46 These tools have undergone extensive validation to ensure they provide estimates similar to those from the full ADVANCE risk equation (Figure 2).

Performance of existing global risk tools for cardiovascular risk estimation in people with diabetes
Two systematic reviews have examined the performance of CVD risk estimation models applicable to people with diabetes.42,47 The most recent and comprehensive review described 45 risk models.42 Of these, 12 were specifically developed for patients with type 2 diabetes (including the ADVANCE model), and 33 were developed in the general population and used diabetes status as a predictor. These models vary greatly in quality and in methodology used to develop them. A number of the prediction models were developed before the advent of novel and more appropriate statistical methods for assessing model performance. Only about one-third of the existing CVD risk tools applicable to diabetics have been externally validated in a population with diabetes.

The discriminative ability both for diabetes-specific CVD prediction models and general population prediction models using diabetes status as a predictor was generally acceptable- to-good (ie, C statistic ≥0.70). However, performance assessment methods used in the validation studies and the discriminative ability in these models varied widely (C statistics: 0.61 to 0.86). The discrimination of prediction models designed for the general population was moderate (C statistics: 0.59 to 0.80) and calibration was generally poor.

The most commonly validated models were the general population– based Framingham cardiovascular risk equations and the diabetes-specific UKPDS risk engines. The Framingham prediction models also showed low-to-acceptable discrimination and poor calibration. Although the discriminative power of UKPDS engines was acceptable, calibration was poor and there was a tendency toward systematic overestimation of risk, particularly in recent cohorts. The models with best external validity were more contemporary, but were validated in other patient populations only once.42 Therefore, more validation studies on the performance of these prediction models in different diabetes populations are needed.


The quest for the best approaches to assess cardiovascular risk and thus prevent vascular complications in individuals with diabetes is a continuing pursuit. It is increasingly clear that diabetes is not a cardiovascular risk equivalent in all circumstances. The CVD risk is not uniformly distributed in individuals with diabetes, but rather follows a gradient. Adequately capturing this gradient depends on the combination of individual risk factors. Global risk assessment appears to be the way forward for managing CVD risk in people with diabetes. Both ADVANCE and subsequent studies have provided evidence that existing popular models derived from older cohorts are less accurate with regard to cardiovascular risk evaluation in a contemporary population with diabetes.42 The recognition of this nonoptimal performance and other limitations of existing models have stimulated efforts to develop new cardiovascular risk models (including the ADVANCE model12) with improved predictive accuracy for people with diabetes. The ADVANCE model is unique in that it was developed from a contemporary multinational cohort of people with diabetes, and successfully validated in another recent multinational cohort of individuals with diabetes. Inclusion of participants from developing countries in the ADVANCE cohort contributes to the potential of the ADVANCE risk model in assisting cardiovascular risk stratification efforts in many settings around the world. _

1. Moons K, Kengne AP, Grobbee DE, et al. Risk predictionmodels: II. External validation, model updating, and impact assessment. Heart. 2012;98(9):691-698.
2. Moons K, Kengne AP, Woodward M, et al. Risk prediction models: I. Development, internal validation, and assessing the incremental value of a new (bio)- marker. Heart. 2012;98(9):683-690.
3. Kengne AP, Patel A, Colagiuri S, et al. The Framingham and UKPDS risk equations do not reliably estimate the probability of cardiovascular events in a large ethnically diverse sample of patients with diabetes: the Action in Diabetes and Vascular Disease: Preterax and Diamicron-MR Controlled Evaluation (ADVANCE) Study. Diabetologia. 2010;53(5):821-831.
4. Echouffo-Tcheugui JB, Kengne AP, Sobngwi E. Cardiovascular risk evaluation tools specific to population with diabetes. Arch Intern Med. 2012;172(6):523-524.
5. Kengne AP, Echouffo-Tcheugui JB, Sobngwi E. Coronary artery calcium for guiding statin treatment. Lancet. [Comment Letter]. 2012;379(9813):312; author reply 312-313.
6. Haffner SM, Lehto S, Ronnemaa T, Pyorala K, Laakso M. Mortality from coronary heart disease in subjects with type 2 diabetes and in nondiabetic subjects with and without priormyocardial infarction. N Engl J Med. 1998;339(4):229-234.
7. Colhoun HM, Betteridge DJ, Durrington PN, et al. Primary prevention of cardiovascular disease with atorvastatin in type 2 diabetes in the Collaborative Atorvastatin Diabetes Study (CARDS): multicentre randomised placebo-controlled trial. Lancet. 2004;364(9435):685-696.
8. Collins R, Armitage J, Parish S, Sleigh P, Peto R. MRC/BHF Heart Protection Study of cholesterol-lowering with simvastatin in 5963 people with diabetes: a randomised placebo-controlled trial. Lancet. 2003;361(9374):2005-2016.
9. Bulugahapitiya U, Siyambalapitiya S, Sithole J, Idris I. Is diabetes a coronary risk equivalent? Systematic review and meta-analysis. Diabet Med. 2009;26(2):142- 148.
10. Gaede P, Lund-Andersen H, Parving HH, Pedersen O. Effect of a multifactorial intervention on mortality in type 2 diabetes. N Engl J Med. 2008;358(6):580- 591.
11. Gaede P, Vedel P, Larsen N, Jensen GV, Parving HH, Pedersen O. Multifactorial intervention and cardiovascular disease in patients with type 2 diabetes. N Engl J Med. 2003;348(5):383-393.
12. Kengne AP, Patel A, Marre M, et al. Contemporary model for cardiovascular risk prediction in people with type 2 diabetes. Eur J Cardiovasc Prev Rehabil. 2011;18(3):393-398.
13. Grundy SM, Cleeman JI, Merz CN, et al. Implications of recent clinical trials for the National Cholesterol Education Program Adult Treatment Panel III guidelines. Circulation. 2004;110(2):227-239.
14. Buse JB, Ginsberg HN, Bakris GL, et al. Primary prevention of cardiovascular diseases in people with diabetes mellitus: a scientific statement from the American Heart Association and the American Diabetes Association. Circulation. 2007;115(1):114-126.
15. IDF Clinical Guidelines Task Force. Global guideline for type 2 diabetes. Brussels: International Diabetes Federation, 2005. http://www.idf.org/webdata/docs/ IDF%20GGT2D.pdf. Accessed October 9, 2010.
16. The Diabetes Control and Complications Trial Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med. 1993; 329(14):977-986.
17. UK Prospective Diabetes Study (UKPDS) Group. Effect of intensive bloodglucose control with metformin on complications in overweight patients with type 2 diabetes (UKPDS 34). Lancet. 1998;352(9131):854-865.
18. UK Prospective Diabetes Study (UKPDS) Group. Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet. 1998; 352(9131):837-853.
19. Patel A, MacMahon S, Chalmers J, et al. Intensive blood glucose control and vascular outcomes in patients with type 2 diabetes. N Engl J Med. 2008;358 (24):2560-2572.
20. Patel A, MacMahon S, Chalmers J, et al. Effects of a fixed combination of perindopril and indapamide on macrovascular and microvascular outcomes in patients with type 2 diabetes mellitus (the ADVANCE trial): a randomised controlled trial. Lancet. 2007;370(9590):829-840.
21. Gonzalez-Clemente JM, Palma S, Arroyo J, et al. Is diabetes mellitus a coronary heart disease equivalent? Results of a meta-analysis of prospective studies [in Spanish]. Rev Esp Cardiol. 2007;60(11):1167-1176.
22. Howard BV, Best LG, Galloway JM, et al. Coronary heart disease risk equivalence in diabetes depends on concomitant risk factors. Diabetes Care. 2006; 29(2):391-397.
23. Sattar N, Preiss D, Murray HM, et al. Statins and risk of incident diabetes: a collaborative meta-analysis of randomised statin trials. Lancet. 2010;375(9716): 735-742.
24. Preiss D, Seshasai SR, Welsh P, et al. Risk of incident diabetes with intensivedose compared with moderate-dose statin therapy: a meta-analysis. JAMA. 2011;305(24):2556-2564.
25. Culver AL, Ockene IS, Balasubramanian R, et al. Statin use and risk of diabetes mellitus in postmenopausal women in the women’s health initiative. Arch Intern Med. 2012;172(2):144-152.
26. Seshasai SR, Wijesuriya S, Sivakumaran R, et al. Effect of aspirin on vascular and nonvascular outcomes: meta-analysis of randomized controlled trials. Arch Intern Med. 2012;172(3):209-216.
27. Lloyd-Jones DM. Cardiovascular risk prediction: basic concepts, current status, and future directions. Circulation. 2010;121(15):1768-1777.
28. Echouffo-Tcheugui JB, Ogunniyi MO, Kengne AP. Estimation of absolute cardiovascular risk in individuals with diabetes mellitus: rationale and approaches. ISRN Cardiol. 2011;2011:242656.
29. Wannamethee SG, Shaper AG, Whincup PH, Lennon L, Sattar N. Impact of diabetes on cardiovascular disease risk and all-cause mortality in older men: influence of age at onset, diabetes duration, and established and novel risk factors. Arch Intern Med. 2011;171(5):404-410.
30. Stevens RJ, Kothari V, Adler AI, Stratton IM. The UKPDS risk engine: a model for the risk of coronary heart disease in type II diabetes (UKPDS 56). Clin Sci (Lond). 2001;101(6):671-679.
31. Kothari V, Stevens RJ, Adler AI, et al. UKPDS 60: risk of stroke in type 2 diabetes estimated by the UK Prospective Diabetes Study risk engine. Stroke. 2002;33(7):1776-1781.
32. Asia Pacific Cohort Studies Collaboration. Systolic blood pressure, diabetes and the risk of cardiovascular diseases in the Asia-Pacific region. J Hypertens. 2007; 25(6):1205-1213.
33. Asia Pacific Cohort Studies Collaboration. Cholesterol, diabetes and major cardiovascular diseases in the Asia-Pacific region. Diabetologia. 2007;50(11):2289- 2297.
34. Asia Pacific Cohort Studies Collaboration. Smoking, diabetes and cardiovascular diseases in men in the Asia-Pacific region. J Diabetes. 2009;1:173-181.
35. Kengne AP, Batty GD, Hamer M, Stamatakis E, Czernichow S. Association of C-reactive protein with cardiovascular disease mortality according to diabetes status: pooled analyses of 25,979 participants from four U.K. prospective cohort studies. Diabetes Care. 2012;35(2):396-403.
36. Coutinho M, Gerstein HC, Wang Y, Yusuf S. The relationship between glucose and incident cardiovascular events. A metaregression analysis of published data from 20 studies of 95,783 individuals followed for 12.4 years. Diabetes Care. 1999;22(2):233-240.
37. Selvin E, Marinopoulos S, Berkenblit G, et al. Meta-analysis: glycosylated hemoglobin and cardiovascular disease in diabetes mellitus. Ann Intern Med. 2004; 141(6):421-431.
38. Miettinen H, Haffner SM, Lehto S, Ronnemaa T, Pyorala K, Laakso M. Retinopathy predicts coronary heart disease events in NIDDM patients. Diabetes Care. 1996;19(12):1445-1448.
39. van Hecke MV, Dekker JM, Stehouwer CD, et al. Diabetic retinopathy is associated with mortality and cardiovascular disease incidence: the EURODIAB prospective complications study. Diabetes Care. 2005;28(6):1383-1389.
40. Targher G, Bertolini L, Tessari R, Zenari L, Arcaro G. Retinopathy predicts future cardiovascular events among type 2 diabetic patients: The Valpolicella Heart Diabetes Study. Diabetes Care. 2006;29(5):1178.
41. Juutilainen A, Lehto S, Ronnemaa T, Pyorala K, Laakso M. Retinopathy predicts cardiovascularmortality in type 2 diabeticmen and women. Diabetes Care. 2007; 30(2):292-299.
42. van Dieren S, Beulens JW, Kengne AP, et al. Prediction models for the risk of cardiovascular disease in patients with type 2 diabetes: a systematic review. Heart. 2012;98(5):360-369.
43. Anderson KM, Odell PM, Wilson PW, Kannel WB. Cardiovascular disease risk profiles. Am Heart J. 1991;121(1 Pt 2):293-298.
44. D’Agostino RB Sr, Vasan RS, Pencina MJ, et al. General cardiovascular risk profile for use in primary care: the Framingham Heart Study. Circulation. 2008;117 (6):743-753.
45. Marre M, Lievre M, Chatellier G, Mann JF, Passa P, Menard J. Effects of low dose ramipril on cardiovascular and renal outcomes in patients with type 2 diabetes and raised excretion of urinary albumin: randomised, double blind, placebo controlled trial (the DIABHYCAR study). BMJ. 2004;328(7438):495.
46. The ADVANCE Collaborative Group. ADVANCE Risk Engine. http://www.ad vanceriskengine.com/index.html. Published 2012. Accessed April 6, 2012.
47. Chamnan P, Simmons RK, Sharp SJ, Griffin SJ, Wareham NJ. Cardiovascular risk assessment scores for people with diabetes: a systematic review. Diabetologia. 2009;24;52(10):2001-2014.

Keywords: ADVANCE risk model; ADVANCE study; cardiovascular; diabetes mellitus; performance; risk prediction