The increase in statin eligibility by the new cholesterol guidelines is mostly driven by the Pooled Cohort Equation (PCE) criterion (≥7.5% 10-year PCE). The impact of replacing the PCE with either the modified Framingham Risk Score (FRS) or the Systematic Coronary Risk Evaluation (SCORE) on assessment of atherosclerotic cardiovascular disease (ASCVD) risk assessment and statin eligibility remains unknown. We assessed the comparative benefits of using the PCE, FRS, and SCORE for ASCVD risk assessment in the Multi-Ethnic Study of Atherosclerosis. Of 6,815 participants, 654 (mean age 61.4 ± 10.3; 47.1% men; 37.1% whites; 27.2% blacks; 22.3% Hispanics; 12.0% Chinese-Americans) were included in analysis. Area under the curve (AUC) and decision curve analysis were used to compare the 3 risk scores. Decision curve analysis is the plot of net benefit versus probability thresholds; net benefit = true positive rate − (false positive rate × weighting factor). Weighting factor = Threshold probability/1 − threshold probability. After a median of 8.6 years, 342 (6.0%) ASCVD events (myocardial infarction, coronary heart disease death, fatal or nonfatal stroke) occurred. All 4 risk scores had acceptable discriminative ability for incident ASCVD events; (AUC [95% CI] PCE: 0.737 [0.713 to 0.762]; FRS: 0.717 [0.691 to 0.743], SCORE (high risk) 0.722 [0.696 to 0.747], and SCORE (low risk): 0.721 [0.696 to 0.746]. At the ASCVD risk threshold recommended for statin eligibility for primary prevention (≥7.5%), the PCE provides the best net benefit. Replacing the PCE with the SCORE (high), SCORE (low) and FRS results in a 2.9%, 8.9%, and 17.1% further increase in statin eligibility. The PCE has the best discrimination and net benefit for primary ASCVD risk assessment in a US-based multiethnic cohort compared with the SCORE or the FRS.
In 2013, the American College of Cardiology (ACC) and the American Heart Association (AHA), at the direction of the National Heart Lung and Blood Institute, developed new guidelines on the treatment of blood cholesterol and risk assessment of atherosclerotic cardiovascular disease (ASCVD). The new ACC/AHA cholesterol guidelines replaced the old National Cholesterol Education Program/Adult Treatment Panel III (ATP III) cholesterol guidelines in the USA. Accompanying this change in cholesterol guidelines was the replacement of the modified Framingham Risk Score (FRS) by the Pooled Cohort Equation (PCE) for ASCVD risk assessment. The change in cholesterol guidelines generated concerns and controversies among researchers and clinicians. One of the major concerns raised was the significant increase in statin eligibility especially for primary prevention of ASCVD which is mostly driven by the PCE risk criterion (PCE ≥7.5% 10-year estimated risk for ASCVD). The modified FRS, used for CVD risk assessment during the era of the National Cholesterol Education Program/ATP III ; the SCORE, used by European Society of Cardiology/European Atherosclerosis Society for CVD risk assessment and the PCE, used by the 2013 ACC/AHA cholesterol guidelines for CVD risk assessment, have similar CVD risk factor constituents although they were developed to predict different cardiovascular disease outcomes. It remains unclear what impact replacing the PCE with either the modified FRS or the SCORE would have on ASCVD risk assessment and the effect it would also have on statin eligibility in the USA population. In this report, we used baseline data from the ongoing Multi-Ethnic Study of Atherosclerosis (MESA) and 10 years of adjudicated ASCVD events to compare the predictive accuracy of the PCE, modified FRS and the SCORE, and their impact on statin eligibility using the ≥7.5% 10-year risk threshold recommended in the new ACC/AHA cholesterol guidelines.
Methods
Details of the MESA have been reported previously. Briefly, from July 2000 to September 2002, a total of 6,814 subjects were recruited at 6 field centers (Baltimore, Maryland; Chicago, Illinois; Forsyth County, North Carolina; Los Angeles, California; New York, New York; and St. Paul, Minnesota). Participants were required to be aged between 45 and 84 years and to have no clinical cardiovascular disease. All participants provided informed consent, and the study protocol was approved by the Institutional Review Boards at each participating institution. For the purpose of this analysis, participants were excluded if they were on cholesterol-reducing medication use at baseline and missing characteristics (n = 1,160), leaving 5,654 subjects for analysis. This was done because the risk scores are recommended to be used for subjects without these factors.
Participant characteristics were collected during the initial MESA visit. Age, gender, and race/ethnicity and a family history of coronary artery disease were self-reported. Smoking was defined as current if they have smoked within the past 6 months. Blood samples were obtained after a 12-hour fast; measurements of total cholesterol and high-density lipoprotein cholesterol were used. Blood pressure was measured for each participant after 5 minutes in the seated position, and systolic measurements were recorded 3 separate times and the mean of the last 2 values was used. The use of antihypertensive medication was self-reported.
We calculated the 10-year risk of incident ASCVD for each individual based on baseline characteristics (age, gender, race/ethnicity, systolic blood pressure, treatment of hypertension, total, high-density lipoprotein cholesterol levels, current smoking), using FRS, SCORE, and PCE. The risk coefficients for SCORE differ by high- and low-risk European populations; thus, we used both SCORE (high-risk countries) and SCORE (low-risk countries) estimates. For Hispanics and Asians, we used white race estimates for the calculation.
The primary outcome was incident ASCVD, composed of fatal and nonfatal myocardial infarction, other fatal and nonfatal coronary heart disease, fatal and nonfatal cerebrovascular disease, and fatal/nonfatal other atherosclerotic disease. Although the risk scores were calibrated for different outcomes, we used the most clinically relevant outcome for all of them which was incident ASCVD. Only first occurrence of ASCVD event was included in the analysis. Participants were followed for incident ASCVD from baseline through December 31, 2012. Follow-up time was defined as the time between the baseline risk score measurement until a diagnosis of ASCVD, loss to follow-up, or end of follow-up (December 31, 2012). At intervals of 9 to 12 months, a telephone interviewer contacted each participant to inquire about all interim hospital admissions, cardiovascular outpatient diagnoses, procedures, and deaths. In addition, MESA occasionally identified medical encounters through cohort clinic visits, participant call-ins, medical record abstractions, or obituaries. Next of kin interviews for out of hospital cardiovascular deaths also were also used. The detailed definitions of these ascertained events are given on http://www.mesa-nhlbi.org/ .
Baseline categorical and continuous variables were compared between statin eligible subjects using chi-square and analysis of variance testing where appropriate. Proportions and means with SDs for continuous and categorical variables were calculated. We computed Harrell’s concordance index (C-index) for FRS, SCORE (high-risk countries), SCORE (low-risk countries), and PCE for incident ASCVD. We then performed decision curve analysis (DCA) using each risk score as a continuous variable. DCA is an analytic instrument to assess net benefit of a diagnostic tool for which there are competing benefits and harms. DCA was performed to compute the net benefit of decision for initiating statin therapy based on 10-year risk assessed using the 4 risk scores at various thresholds.
Net Benefit = True positive rate − ( false positive rate × weighting factor )
Weighting factor = Threshold probability/(1 − threshold probability) and threshold probability is 10 years. ASCVD risk above which a subject becomes eligible for statin therapy.
The desirable outcome, or “benefit,” was defined with the thought that initiation of statin therapy might have prevented incident ASCVD during follow-up period, whereas the undesirable outcome, or “harm,” was defined with the thought that initiation of statin therapy might have not prevented incident ASCVD in subjects who developed incident ASCVD during the follow-up period. The DCA generates a graph of net benefit as a function of a threshold probability (p t ) of 10-year risk of incident ASCVD at which an individual considers the potential benefit and harm of statin therapy to be equivalent. The net benefit was measured as the rate that incorporating the decision guide of interest (such as initiating statin therapy) would lead to additional beneficial decision to treat high-risk subjects without causing any additional harmful decision to overtreating low-risk subjects. A p value of <0.05 was considered significant. DCA was performed using the R statistics package. Analyses were performed using PASW, version 18.0 (IBM Inc., New York, New York).
Results
A total of 5,654 (mean age 61.4 ± 10.3; 47.1% men; 37.1% whites; 27.2% blacks; 22.3% Hispanics; 12.0% Chinese-Americans) were included in the final analysis. After a median follow-up of 8.5 years (interquartile range 7.6 to 8.6 years), there were 342 (6.0%) incident ASCVD events. Table 1 lists the baseline characteristics of the participants included in this analysis. As shown in Figure 1 , the PCE had the best area under the curve (AUC), whereas the FRS had the least AUC for discrimination of ASCVD events in this cohort. The AUCs of both high-risk and low-risk countries SCORE were not significantly different, whereas AUC of FRS was significantly different from PCE AUC (p = 0.04).
| Characteristics | All cohort |
|---|---|
| Age (years) | 60 ± 10 |
| Male | 47% |
| White | 39% |
| Black | 27% |
| Asian | 12% |
| Hispanic | 22% |
| Hypertension | 36% |
| Anti-hypertensive use | 25% |
| Smoker | 50% |
| Family history of coronary artery disease | 32% |
| Systolic blood pressure (mm Hg) | 124 ± 21 |
| Total Cholesterol (mg/dL) | 196 ± 35 |
| Low density lipoprotein (mg/dL) | 120 ± 31 |
| High density lipoprotein (mg/dL) | 51 ± 15 |
| Glomerular filtration rate (ml/min) | 83 ± 18 |
| Body mass index (kg/m 2 ) | 28 ± 5 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
