International Journal of Cardiology Volume 176, Issue 3, 20 October 2014, Pages 1154–1155
 

Comparison of pooled cohort risk equations and Framingham risk score for metabolic syndrome in a Korean community-based population

Jaewon Oh, Seok-Min Kang, Namki Hong,Jong-Chan Youn,Sungha Park,Sang-Hak Lee,Donghoon Choi

 

Keyword

• Risk assessment;


• Metabolic syndrome;


• Coronary heart disease

Previous risk assessment guidelines, such as the Adult Treatment Panel III (ATP-III), used a modified Framingham risk score (FRS) to set thresholds for lipid treatment based on one's predicted 10-year risk (10YR) for myocardial infarction or coronary heart disease (CHD) death. In 2013, the American Heart Association/American College of Cardiology guideline reported a new pooled cohort risk equation (PCE) designed to assess the 10YR of atherosclerotic cardiovascular disease (ASCVD) [1]. However, there have been some arguments about miscalculations of the new equation such as risk overestimation and limitation of performance for diverse ethnics [2] and [3]. For instance, the new PCE systematically overestimated observed risks by 75–150%, roughly doubling the actual observed risk [3]. And recent study showed that FRS overestimated the risk of CHD in the Korean population where the CHD incidence is low [4]. So we compared the performance of new PCE and old FRS in terms of metabolic syndrome (MetS) in a Korean community-based population. This is the first report about the performance of PCE in an Asian population as far as we know.

We analyzed the cross-sectional data of 12,700 participants from the 2009–2010 Korea National Health and Nutritional Examination Survey (KNHANES) [5]. According to the adaptation rule of PCE suggested in the guideline, we excluded the subjects with 1) ages under 20 & over 80 (n = 4382) and 2) a history of ASCVD (n = 496). PCE and FRS were calculated with the instructions of the guideline [1]. MetS was defined in accordance with the revised National Cholesterol Education Program Adult Treatment Panel (ATP) III criteria.

In this population (n = 7,822), mean 10YRs of PCE and FRS were 9.3 ± 10.4% and 14.5 ± 14.2%, respectively, and the prevalence of MetS was 36.8% (n = 2882). In Pearson correlation analysis, 10YRs by two equations were significantly correlated (r = 0.919, p < 0.0001). The prevalence of MetS increased gradually according to the 10YR categorical group (from 10YR < 2.5% to 10YR > 20%) using both equations (PCE, 16.4% to 60.9%; FRS, 5.1% to 57.9%; p for trend < 0.001 for all, Fig. 1). The definitions of the intermediate risk group by two equations were different such as 5% ≤ 10YR < 7.5% in PCE but 10% ≤ 10YR < 20% in FRS, so the prevalence of low, intermediate and high risk groups 47.9% (n = 3749), 10.6% (n = 830), and 41.5% (n = 3243) in PCE and 50.5% (n = 3952), 25.1% (n = 1959), and 24.4% (n = 1911) in FRS, respectively. When using PCE, the high risk group, in other words, who needs a statin therapy was increased (about 1.7 times), consistent with a recent study in the United States and in Europe [6] and [7].
 In Bland–Altman analysis, 10YR by PCE was lower than that by FRS (median difference; 3.7%, 95% confidence interval; − 1.67%–17.63%, Fig. 2). Recent validation study reported that FRS overestimated the risk of CHD in Korean. In line of these backgrounds, we hypothesized that 10YR by PCE could be more reliable than 10YR by FRS. So we compared the performance of two equations using two statistical methods. Firstly, the receiver-operating characteristic (ROC) analyses were performed to compare the diagnostic performance of the 2 equations for MetS. The AUC for 10YR by PCE itself was lower than that by FRS (0.709 vs. 0.728, p < 0.0001), and the AUC for 10YR categorical group by PCE was also lower than that by FRS (0.694 vs. 0.722, p < 0.0001). Secondly, we had a net reclassification improvement (NRI) analysis for comparing the discriminating power of the two equations. The continuous NRI (PCE over FRS) by 10YR itself and the risk group was 0.126 (95% CI, − 0.017–0.189) and 0.235 (95% CI, 0.139–0.277), respectively. A recent Canadian study also showed that PCE did not show a marked propensity to reclassify low FRS to higher risk levels [8], consistent with our finding.
 

Our study does not go without limitations. Firstly, we could not compare the hard clinical outcomes such as cardiovascular mortality and myocardial infarction using the two equations, because the follow-up data was not available in KNHANES. It has been one of the biggest limitations of this national representative survey data. Secondly, we could not analyze the additive prognostic value of a high sensitive C-reactive protein (hsCRP), which is also recommended by the 2013 guideline as a screening test for ASCVD. The biomarkers such as hsCRP were not available in our data set. In addition, a recent study showed that hsCRP concentrations are substantially lower in Koreans than that reported in the White populations, so the cutoff value of hsCRP should be specified for Asians including Koreans [9]. Thirdly, it will be interesting to study whether there are any additive prognostic values of imaging screening tests (e.g. coronary calcium score by computed tomographic scan, intima–media thickness by carotid ultrasonography) to PCE for individuals with an intermediate risk.

In conclusion, the new PCE by the 2013 guideline did not improve the net reclassification for MetS risk stratification in the Korean community-based population. Further validation study which compares the clinical outcomes using two equations should be warranted before using the new PCE in a real world setting in an Asian population.

Conflict of interest

None declared.

Acknowledgments

This work was supported by a grant of the Korean Health Technology R&D Project, Ministry of Health & Welfare, Republic of Korea (HI13C0715).