Physicians' Academy for Cardiovascular Education

Novel CV risk score can be adapted according to age and specific situation in a country

Hajifathalian K et al., Lancet Diabetes Endocrinol 2015

A novel risk score to predict cardiovascular disease risk in national populations (Globorisk): a pooled analysis of prospective cohorts and health examination surveys


Hajifathalian K, Ueda P, Lu Y et al.
Lancet Diabetes Endocrinol 2015 Online: 25 March. DOI: http://dx.doi.org/10.1016/S2213-8587(15)00081-9
 

Background

Consensus exists that treatment for cardiometabolic risk factors such as blood pressure and cholesterol should be guided by disease risk, rather than on the individual risk factors [1-3], although the appropriate threshold for treatment is still topic of debate. Risk-based multidrug treatment and counselling could be a cost-effective means to reduce the burden of non-communicable diseases worldwide [4].
A risk score estimates a person’s risk of CV disease during a specific period based on their risk factors and the average CV disease risk in the population, and is often build up from hazard ratios (HRs) of individual CV risk factors. HRs for major CV disease risk factors are thought to be similar in Western and Asian populations, and over time within a population. CV disease risk, however, substantially differs between populations and over time.
Thus, risk prediction equations cannot be satisfactorily applied to other populations of even in the same populations a while after the risk score was developed [5,6]. Recalibration to the average risk factor level of the target population can be performed to fix this problem, although this has yielded mixed results.
The current paper describes a risk prediction equation that can be recalibrated and updated for use in different countries with routinely available information. Individual-level data of persons aged 40 years or older without a history of coronary heart disease or stroke from 8 prospective cohorts were pooled (including cohorts in Honolulu and Puerto Rico). Three additional non-USA based cohorts (Scotland, Tehran/Iran, Australia) were used for validation of the risk prediction equation.
 

Main results

  • The developed risk score for fatal CV disease performed well in internal validation (median C statistic was 73.5%, with a range of 60-78% when data from different cohorts were omitted and used for validation).
  • In external validation, the Scottish Heart Health Extended Cohort gave a C statistic of 74%, (calibration χ2: 12.0), the Tehran Lipid and Glucose study 83% (calibration χ2: 12.9) and the Australian Diabetes, Obesity and Lifestyle cohort 84% (calibration χ2: 44.3, mainly due to the highest decile of predicted risk).
  • The estimated 10-year fatal CV disease risk score, at any age and risk factor level, varied substantially between various national populations, with the lowest risk in Japan, South Korea, Spain, Denmark and England, and the highest in China and Mexico.
  • Total-to-fatal CV disease ratio tended to decrease with age, implying that more fatal events occur at older ages. Total-to-fatal CV disease risk ratios were expected to be decreased in low-income and middle-income countries. Indeed substantial differences in fatal CV disease risk distributions were seen across countries.
  • The Globorisk risk prediction tool will become available online.

Conclusion

The developed risk prediction equation and risk charts that can be generated with it can be used to identify individual patients with a high risk of CV disease in different countries, and to estimate the number of such people in a population. The latter allows measuring progress towards public health goals.
Using the risk score suggested a substantially increased prevalence of people at high CV risk in low-income and middle-income countries, as compared with high-income countries.
The equation allows for variation of the age and sex pattern of CV disease risk across populations and over time. This risk prediction equation facilitates global application of risk stratification.  
 

Editorial comment [7]

While most CV disease prediction models are developed in North American or European countries, often they are not validated for predictive accuracy in individuals outside the population they were developed for. The authors note that CV disease burden is also rapidly increasing in low-income and middle-income countries, thus adaptation of risk models for these countries is needed.
“The Globorisk model was developed such that country specific recalibration was feasible with few data. However, if individual participant-level data from the target setting are available, existing models can often be similarly recalibrated, which would allow for head-to-head comparisons between models per country. Such direct comparisons will further help health-care professionals and policy-makers in these countries to decide which model to promote, which puts the findings of Hajifathalian and colleagues in an even broader perspective.” (…)
“The country-specific predictions for estimated 10 year cardiovascular disease burden are striking, particularly the large proportion of high-risk individuals in China, Mexico, Czech Republic, and Iran. A next step would be to quantify the positive effects on a population level if the Globorisk model and subsequent risk based preventative management were used in these countries.”
 
Find this article online at The Lancet Diabetes and Endocrinology
 

References

1 Jackson R, Lawes CM, Bennett DA, et al. Treatment with drugs to lower blood pressure and blood cholesterol based on an individual’s absolute cardiovascular risk. Lancet 2005; 365: 434–41.
2 JBS3 Board. Joint British Societies’ consensus recommendations for the prevention of cardiovascular disease. Heart 2014; 100 (suppl 2): ii1–67.
3 WHO. Prevention of cardiovascular disease: guidelines for assessment and management of cardiovascular risk. Geneva: World Health Organization, 2007.
4 Lim SS, Gaziano TA, Gakidou E, et al. Prevention of cardiovascular disease in high-risk individuals in low-income and middle-income countries: health effects and costs. Lancet 2007; 370: 2054–62.
5 Cook NR, Paynter NP, Eaton CB, et al. Comparison of the Framingham and Reynolds Risk scores for global cardiovascular risk prediction in the multiethnic Women’s Health Initiative. Circulation 2012; 125: 1748–56.
6 Neuhauser HK, Ellert U, Kurth BM. A comparison of Framingham and SCORE-based cardiovascular risk estimates in participants of the German National Health Interview and Examination Survey 1998. Eur J Cardiovasc Prev Rehabil 2005; 12: 442–50.
7 Moons KGM, Schuit E. Prediction of cardiovascular disease worldwide. Lancet Diabetes Endocrinol 2015 Published Online March 26 date, 2015 http://dx.doi.org/10.1016/
S2213-8587(15)00002-9