How to calculate vascular age with the SCORE project scales: a new method of cardiovascular risk evaluation

Abstract

Introduction

According to the WHO, cardiovascular diseases (CVD) are the leading cause of death on the European continent,¹ accounting for 44% of all deaths in men and 57% in women in 2004. CVD are responsible for 4.7 million deaths annually in Europe,² with over 2 million deaths in the European Union in 2008.³ CVD are also responsible for a very important health-care problem in terms of both direct economic costs as well as social repercussions resulting from incapacity, causing direct costs of over 100 000 million Euros yearly in the European Union.³

As a cause of CVD, cardiovascular risk factors have a high prevalence in our setting, and whilst some factors have been reduced, such as smoking, the prevalence of others, such as diabetes and obesity, is increasing.³ Additionally, the degree of control of the various cardiovascular risk factors is very low, in both the primary care and the specialized care settings.⁴

Clinical care guidelines for the management of patients with cardiovascular risk factors currently require the measurement of cardiovascular risk as the starting point to establish therapeutic aims and treatment strategies.^5,6 The Third and Fourth Joint European Task Forces on Cardiovascular Prevention established the SCORE system⁷ as the instrument to quantify the absolute cardiovascular risk.^6,8 The level of risk is used to determine the objectives for control of the various risk factors and the need or otherwise for pharmacological intervention, in addition to lifestyle measures.

Therapeutic strategies based on cardiovascular risk afford the same or greater benefits than strategies based on the isolated control of each individual risk factor, as well as being cheaper.⁹

One of the limitations of any system that estimates risk, and in particular the SCORE system, is that when it is applied to middle-aged persons to calculate the absolute risk, this is not high even though their relative risk may be high. This situation can lead to no preventive or therapeutic decisions being taken in these persons. Accordingly, the Third Joint European Task Force⁸ recommended extrapolating the risk for the age of 60 years, though the literal application of this recommendation may result in pharmacological overtreatment in these persons.⁶ The Fourth Joint Task Force recommended the use of the relative risk together with the absolute risk, although no cut point was established above which the relative risk is considered high.⁶ Another alternative to solve this limitation is the use of percentiles of cardiovascular risk to place each person within the context of risk for persons of the same age and gender.¹⁰

Among other physician-patient aspects, the Fourth Task Force established that preventive measures should be based on a patient-centred approach, where the therapeutic decisions are taken in collaboration with the patient in order for the patient to assume them personally and where an evaluation is made of how to communicate the risk.⁶

In addition to the SCORE system, based on a European population, many other methods exist to quantify cardiovascular risk, most originating with the Framingham Heart Study (FHS),¹¹ started in 1948, and based on an American population, whose overall risk is higher than that of Europeans. In 2008, as a result of the Framingham study, cardiovascular risk tables were published,¹² which incorporated a new concept: the age of the heart or the vascular age, i.e. the age of the vascular system of a patient with different cardiovascular risk factors. This age is calculated as the age a person would be with the same calculated cardiovascular risk but whose risk factors were all within normal ranges.

Absolute cardiovascular risk is a statistical and epidemiological concept that many patients may find difficult to understand. However, the concept of vascular age as the age of the arteries is a concept related to risk that might be more easily understood by all patients. The use of vascular age may thus help the physician explain the risk status of the patient and therefore improve patient compliance with therapeutic decisions.

The aim of this study was to determine the vascular age calculated from a different scale to that of the original 2008 Framingham scale, i.e. with the two SCORE project scales (for high- and low-risk European countries) and to determine whether vascular age is similar with the two SCORE scales.

Methods

Vascular age was calculated according to the definition of D'Agostino in the tables from the 2008 FHS,¹² who defined the vascular age of a person as the age a person would be with the same calculated cardiovascular risk but whose risk factors were all within normal ranges, i.e. with a risk solely due to age and gender. The risk factors considered were age, gender, smoking, total cholesterol (T-Chol) level, systolic blood pressure (SBP) and diabetes.

The FHS defines 10-year absolute CVD risk as the risk at 10 years of suffering a CVD, defined as a composite of coronary heart disease (coronary death, myocardial infarction, coronary insufficiency and angina), cerebrovascular events (including ischaemic stroke, haemorrhagic stroke and transient ischaemic attack), peripheral artery disease (intermittent claudication) and heart failure. To calculate this CVD risk, FHS evaluated 8941 participants (4522 women) from original and offspring cohorts with a maximum follow-up period of 12 years.

Unlike the FHS, the SCORE project calculates 10-year fatal CVD risk as the probability of cardiovascular death. With this aim, the SCORE project assembled a pool of databases from 12 European cohort studies (Finland, Russia, Norway, British region, Scotland, Denmark, Sweden and Germany [for high-risk countries], and Belgium, Italy, France and Spain [for low-risk countries]), mainly relating to the general population. The total involved 205 178 persons (88 080 women) representing 2.7 million person-years of follow-up. There were 7934 cardiovascular deaths, of which 5652 were deaths from coronary heart disease. The risk factors considered by the SCORE project were age, gender, smoking, and T-Chol and SBP levels.

We calculated vascular age from SCORE project equations for different situations of the combination of age, gender, smoking, and T-Chol and SBP levels (as can be seen in the SCORE project tables). We considered 120 mmHg as a normal pressure because it is the limit between normal and optimal SBP according to the 2007 Guidelines for the Management of Arterial Hypertension by the European Societies⁵ and the Fourth Joint Task Force on Cardiovascular Disease Prevention.⁶ The normal level of T-Chol was established at 5 mmol/L (190 mg/dL) as this is the high limit of normality established by the Fourth Joint European Task Force.

Data on diabetes were not collected uniformly in the SCORE study cohorts, so we decided not to include a dichotomous diabetes variable in the risk equations and suggest that among persons with clinically established type 2 diabetes, the increase in CVD risk is at least 2-fold in men and even higher, as much as 4-fold, in women.⁷

The risk of fatal CVD was calculated according to the SCORE project with the risk factors controlled for each combination of age and gender. The mathematical equations used to calculate the risk were the originally published equations⁷ and the risk was calculated for low-risk countries and high-risk countries. The calculations were extended beyond the age of 65 years to cover the highest risks in the fatal CVD tables of the SCORE project.

The data on vascular age are presented as tables in order to be able to use them in combination with the risk tables of fatal CVD. The data are also presented as charts relating vascular age with cardiovascular risk. Colour tables were also created for vascular age, similar to the SCORE tables, to be able to use them directly with the data from each patient, such that with the data on cardiovascular risk factors we can calculate the vascular age directly.

To determine the degree of agreement between vascular age calculated with the two SCORE scales (for high- and low-risk countries), the difference in vascular age between both scales was calculated for each situation of age, gender, smoking, T-Chol and SBP, as well as the intraclass correlation coefficient. Finally, the regression line was calculated between the two vascular ages. The statistical calculations were made with SPSS 15.0.

Results

Table 1 shows the vascular age corresponding to each absolute risk value calculated with the risk equations for high-risk countries, for both men and women. Table 2 shows the same information for low-risk countries.

The data from each table were used to create Figures 1 and 2, which clearly show how vascular age increases with the increase in absolute risk, for both men and women and in both high- and low-cardiovascular-risk countries. The figures also show that the risk did not have a linear relation with age, but rather that the increase in risk was greater the older the person, with an almost exponential relation.

Figure 1

Vascular age in men and women according to the absolute risk of fatal cardiovascular disease in the SCORE project for high-risk countries.

Open in new tabDownload slide

Figure 2

Vascular age in men and women according to the absolute risk of fatal cardiovascular disease in the SCORE project for low-risk countries.

Open in new tabDownload slide

These charts can be used together with the classical colour tables for cardiovascular risk from the SCORE project. Figure 3 shows an example of this possibility: if we consider a non-diabetic 50-year-old man living in a low-risk country who smokes, has a SBP of 160 mmHg and a T-Chol level of 7 mmol/L, we can calculate his absolute risk to be 4%. With the new figure we can see that this corresponds to a vascular age of 69 years (Figure 3).

Figure 3

Example of the use of the vascular age chart for a non-diabetic 50-year-old male smoker with systolic blood pressure of 160 mmHg and total cholesterol of 7 mmol/L. Absolute risk: 4%; vascular age: 69 years. Tables for low-risk countries (modified with permission from Conroy RM⁷).

Open in new tabDownload slide

Figures 4 and 5, for high- and low-risk countries, show the vascular age using the format of the original absolute risk tables, in which each box corresponds to a particular gender, age, smoking status, SBP level and T-Chol concentration. The vascular age calculations were extended beyond 65 years of age, thereby covering all the absolute risk values for men and women in the SCORE tables for high- and low-risk countries. In these new figures, the colour of each box corresponds to the original absolute risk and the number of each box corresponds to vascular age, grouping together in the same table the information for absolute risk and vascular age.

Figure 4

Colour table of vascular age. Each box contains a number corresponding to vascular age and a colour corresponding to the overall absolute risk of fatal cardiovascular disease according to the SCORE project for high-risk countries. SBP, systolic blood pressure.

Open in new tabDownload slide

Figure 5

Colour table of vascular age. Each box contains a number corresponding to vascular age and a colour corresponding to the overall absolute risk of fatal cardiovascular disease according to the SCORE project for low-risk countries. SBP, systolic blood pressure.

Open in new tabDownload slide

To analyze the differences in vascular age for the same T-Chol concentration, systolic blood pressure, age, gender and smoking status between the tables for the high- and the low-risk countries, we provide a new table with the differences in the corresponding boxes of the two tables for vascular age (Figure 6). Of the 400 boxes, 347 (86.75%) differed by 1 year or less, 47 (11.75%) by 2 years and 6 (1.5%) by 3 years. The intraclass correlation coefficient was 0.997.

Figure 6

Differences in calculated vascular age between the two vascular risk models of the SCORE project for high- and low-risk countries. SBP, systolic blood pressure.

Open in new tabDownload slide

Figure 7 shows the regression line analysis. The Pearson correlation coefficient was 0.999, with a line determined by the equation: (vascular age in low-risk countries) = 1.496 + 0.967 × (vascular age in high-risk countries).

Figure 7

Correlation between vascular age calculated with the equations for high-risk countries (VAhigh) and for low-risk countries (VAlow).

Open in new tabDownload slide

Discussion

In this study, we present the calculation of vascular age with a different cardiovascular risk evaluation scale from the original scale that gave rise to the concept, in such a way as to be able to extend the use of vascular age together with the scale most used in Europe. A review of the literature shows this to be the first time that vascular age has been extended for use with other scales apart from the original 2008 Framingham model,¹² and its use with the SCORE project is therefore new.

Cardiovascular risk evaluation is an imprecise tool, not only because the result is a higher or lower populational probability of presenting a cardiovascular event, but also no adequate agreement exists between the various different systems for stratifying or calculating the risk.^13,14 This lack of agreement is due, among other reasons, to the various underlying mathematical models used to evaluate cardiovascular risk, e.g. the Cox proportional hazards model in the Framingham study or the Weibull model in the SCORE project, the different event-defining variables (coronary heart disease in most Framingham-based systems, cardiovascular death in the SCORE project), different variables introduced into the models and different baseline study populations used to calculate the scores. This last factor is of particular importance, because it means that application of a risk evaluation system to a different population from the original study population requires recalibrating the model,¹⁵ or, as with the SCORE project, starting from two models, one for countries with a high cardiovascular risk and another for countries with a low cardiovascular risk.

An important limitation of the use of the SCORE system to evaluate the risk concerns the effect of age. Middle-aged persons do not have a high risk even though their risk factors may be high, whereas older persons have a high risk with just slight increases in their risk factors. A 40-year-old man, even if he smokes, has a SBP of 180 mmHg and a T-Chol of 8 mmol/L, still only has an overall risk of cardiovascular death of less than 5%, whether he lives in a high-risk country (4%) or low-risk country (2%). A 65-year-old man from a high-risk country, however, who does not smoke, has a SBP of 140 mmHg and a T-Chol of 4 mmol/L nevertheless has a risk of 6%, which is high. In order to palliate this effect in middle-aged persons the use of the relative risk has been suggested,⁶ though no consensus has yet been reached about a cut point to define the relative risk as high. Another alternative is the application of risk percentiles,¹⁰ which places each person within the context of the risk for persons of similar age and gender, thereby enabling identification of those persons at greater risk for their age, at the same time as it improves the agreement between risk calculation systems.¹⁶

Another limitation of the use of cardiovascular risk is that it is not clear that all patients understand what a high risk represents in terms of a high likelihood of having a cardiovascular event, which makes it more difficult to communicate this risk. Patients tend to respond to the risk more according to their emotions than facts, overestimating low risks and underestimating common risks.¹⁷ CVD are very common, and patients may underestimate the information received concerning their own cardiovascular risk. Although patients are aware that lifestyle modification may considerably reduce their cardiovascular risk, they still maintain unhealthy lifestyles. For instance, four of every five persons in North America lead lifestyles that place them at an increased risk for coronary and cerebral ischaemic events.¹⁸ Identifying a risk as ‘natural’ is common, which may explain why patients fail to take sufficiently seriously their own cardiovascular risk, which they may partly be able to control, whereas they worry about other ‘unnatural’ risks, which they cannot control, even though the risk may be almost nil, like living near electricity power stations or mobile telephone antennas. Patients also respond very differently according to how risk is presented, for instance relative reduction in risk, absolute reduction or number necessary to treat.¹⁹

The use of vascular age with the data presented here may be a very useful tool in the management of patients with cardiovascular risk factors. The concept of vascular age may be a more reasonable concept for the patient than that of risk, and the patient may thus be able to understand what a particular risk means in terms of life. It therefore represents a valuable instrument to enable patients to assimilate their situation and thus be more likely to comply with therapeutic measures. The 40-year-old male smoker with high blood pressure and high cholesterol, whose risk was lower than 5%, has a vascular age of 63 years. This patient can thus understand that despite not having a high absolute risk, his risk factors make him 23 years older from the vascular point of view. Although the aim of this study was not to show that the concept of vascular age is easier to understand than absolute or relative risk, age would seem to be a more natural concept to comprehend than risk.

Notably, the agreement in vascular age between the two SCORE project scales was almost total. Whichever scale we use, the result in terms of vascular age is practically the same. This agreement permits the use of either scale in all European countries. This finding is completely new. The absolute risk differs greatly according to whether we use the SCORE tables for high- or low-risk countries,⁷ but vascular age is the same. This very high level of agreement was demonstrated by an extraordinarily high intraclass correlation coefficient of 0.997. The great majority, 86.75%, of the different situations of the combination of age, gender, smoking, T-Chol level and SBP differed by no more than 1 year when vascular age was compared between the two SCORE scales. The most discordant values were found in the boxes corresponding to the highest vascular age. The 1.5% of cases where vascular age differed by 3 years showed vascular age to be 30 or nearly 40 years greater than biological age. Thus, the clinical importance of this difference is low.

Analysis according to gender showed a greater agreement for women, as in no case there was a difference of 3 years between the vascular ages calculated with the two scales. In men the agreement was still very high and the differences of 3 years were found in situations where the SBP and the T-Chol level were both high, with a T-Chol level of 8 mmol/L or SBP of 180 mmHg, situations associated with a high cardiovascular risk.

The presentation of vascular age in the form of a coloured chart, similar to that presenting the absolute risk in the SCORE project,⁷ besides using a conversion table between absolute risk and vascular age, enables both the physician and the patient to evaluate the absolute risk and vascular age rapidly.

To obtain all vascular ages we used the equations of the SCORE project beyond 65 years, which is the maximum biological age in the SCORE tables. This may therefore represent an over- or under-estimation of cardiovascular risk. Nevertheless, this calculation was not done to determine the risk in persons older than 65 years of age, but rather to determine the vascular age of persons up to the age of 65 years with a high cardiovascular risk. Even if an error occurred in the evaluation of risk with this group of persons, the main usefulness of the vascular age tables is to help the patients understand that their particular cardiovascular risk factors make their vascular age much higher than their actual biological age, thus retaining their invaluable benefit in patient care.

In summary, we can state that we are able to obtain much more information from the calculation of the absolute risk with the SCORE project tables. The application of the concept of vascular age provides a new use for the risk tables that might be more practical for both the patient and the physician, and its use can even be encouraged among middle-aged persons, for whom there were no satisfactory solutions for the problems associated with the evaluation of absolute risk. The second main finding of this study is that the agreement in vascular age between the two SCORE project scales was almost total. Vascular age may be a practical instrument for patient health education and favour patient compliance concerning cardiovascular status.

Conflict of interest: none declared.

References