Importance: Cardiovascular risk assessment employs traditional risk factors to identify individuals who may benefit from primary prevention therapies. Incorporating nontraditional risk factors may improve traditional multivariate risk assessment.
Objective: To systematically review evidence for the use of nontraditional risk factors—ankle-brachial index (ABI), high-sensitivity C-reactive protein (hsCRP), and coronary artery calcium (CAC)—in asymptomatic adults without known cardiovascular disease (CVD). Five key questions address: clinical impact of nontraditional risk factor assessment versus traditional risk factor assessment with Framingham Risk Score (FRS) or Pooled Cohort Equations (PCE) (KQ1), performance of nontraditional risk factor assessment added to the FRS or PCE (KQ2), harms of nontraditional risk factor assessment (KQ3), and benefits (KQ4) and harms (KQ5) of nontraditional risk factor-guided therapy. The United States Preventive Services Task Force (USPSTF) will use this review to update prior recommendations on the use of nontraditional risk factors and the use of CVD risk assessment with the ABI.
Data Sources: MEDLINE, PubMed, and Cochrane Collaboration Registry of Controlled Trials through May 22, 2017, to update existing systematic reviews supporting the previous USPSTF recommendations.
Study Selection: We screened 22,707 abstracts and 483 full-text articles against a priori inclusion criteria. For KQ1 and KQ4 we limited studies to trials reporting patient health outcomes. For KQ2 we included risk prediction studies comparing a base model with traditional risk factors (the FRS or PCE) to extended models also including one of the three nontraditional risk factors (ABI, hsCRP, CAC) predicting CHD or CVD outcomes. For KQ3 and KQ5 we broadly included any study design examining harms of nontraditional risk assessment or nontraditional risk factor-guided therapy. All KQs were limited to studies of asymptomatic populations that were conducted in developed nations and published in the English language.
Data Extraction and Synthesis: Two investigators independently and critically appraised each article that met inclusion criteria using USPSTF’s design-specific criteria, supplemented by the Checklist for Critical Appraisal and Data Extraction for Systematic Review of Prediction Modelling Studies (CHARMS) for risk prediction studies. Poor-quality studies were excluded. Data from fair- and good-quality trials were abstracted into standardized evidence tools in DistillerSR, with all data double-checked by a second reviewer for accuracy. Due to the limited number of included studies and/or clinical heterogeneity of included studies, we did not conduct meta-analyses. We graded the strength of the overall body of evidence for each KQ.
Main Outcomes and Measures: For KQ1 and KQ4, outcomes included fatal and nonfatal CVD events (e.g., myocardial infarction [MI], cerebrovascular accident [CVA]) and all-cause mortality. For KQ2, outcomes included any measure of calibration (e.g., calibration plot, Hosmer-Lemeshow test) or overall performance (e.g., likelihood ratio tests, R2), discrimination (e.g., c-statistic/area under the curve [AUC]), or reclassification (e.g., net reclassification index [NRI]). For KQ3, outcomes comprised any harms, including radiation exposure due to CT imaging for CAC and downstream health care utilization. For KQ5, outcomes included any serious adverse event as defined by the included study.
Results: We included a total of 43 unique studies reported in 54 publications (some studies were included for multiple KQs): 1 study for KQ1, 33 studies for KQ2, 8 studies for KQ3, 4 studies for KQ4, and 3 studies for KQ5.
KQ1: One fair-quality RCT (n=2,137), primarily designed to assess the impact of CAC on CVD risk factors and downstream testing, reported health outcomes and found no statistically significant differences in CVD events between CAC score and control groups at 4 years. This study was not adequately powered for CVD outcomes.
KQ2: Ten studies (n=81,590) evaluated ABI, 25 studies (n=269,449) evaluated hsCRP, and 19 studies (n=69,720) evaluated CAC. Only four studies evaluated nontraditional risk factors in addition to the PCE; the rest used a base model of FRS. Overall, limited data suggest all three nontraditional risk factors can improve calibration, but the clinical impact of this change in calibration is uncertain due to the lack of reporting of preferred measures. We have more data to inform the change in discrimination and risk reclassification when adding ABI, hsCRP, or CAC to traditional cardiovascular risk assessment.
ABI: One large, individual-participant data (IPD) meta-analysis including 18 different cohorts demonstrated that ABI can improve discrimination and reclassification in women to predict hard CHD events when added to a published coefficient FRS model, with a c-statistic change of 0.112 and net reclassification index (NRI) of 0.096. This incremental improvement for women is most likely due to poorer discrimination of the base model in women, compared to men.
hsCRP: Results for hsCRP are mixed. Studies using published coefficients for FRS demonstrate that hsCRP can improve discrimination, but results are inconsistent. One large IPD meta-analysis, a model development study that included 38 different cohorts, demonstrated that hsCRP only had very small improvement on discrimination. Results for reclassification were similar and best evidence suggests an overall NRI of less than 0.02.
CAC: Based on a smaller body of evidence, CAC consistently appears to improve discrimination and reclassification in both published coefficient and model development studies; NRIs ranged from 0.084 to 0.35.
KQ3: No studies address the harms of ABI or hsCRP. Four studies (n=11,473) demonstrated that radiation exposure from CT imaging for CAC is low. Two studies (n=1,619) found no evidence for adverse psychological health outcomes for screening CAC. Two studies (n=11,364) found no evidence that CAC paradoxically increases CVD events. Three studies (n=13,204) found mixed results for CAC on downstream health care utilization. Best evidence suggests no overall increase in cardiac imaging or revascularization; however, this RCT may have limited applicability to real-world practice. One large retrospective study using Medicare claims data found an association for higher utilization compared to hsCRP or lipid screening.
KQ4: No trials directly compared treatment guided by nontraditional risk factors when added to traditional cardiovascular risk assessment; however, we included studies in which preventive therapies were guided by the use of nontraditional risk factors. Two RCTs (n=4,626) found no benefit for ABI-guided low-dose aspirin on CVD outcomes or all-cause mortality at approximately 7 to 8 years of followup. One RCT (n=17,802) found a benefit for hsCRP-guided, high-intensity statin on CVD outcomes and all-cause mortality at 1.9 years of followup. One RCT (n=1,005) found no benefit for CAC-guided moderate-intensity statin at approximately 4 years, but the study was inadequately powered to detect a benefit for CVD outcomes.
KQ5: Low-dose aspirin in the two RCTs (n=4,626) included for KQ4 did not result in increased major bleeding events. High-intensity statin in one RCT (n=17,802) included in KQ4 was associated with an increase in incident diabetes but not with other serious adverse events.
Conclusions and Relevance: There is no direct evidence from adequately powered clinical impact trials comparing traditional cardiovascular risk assessment to risk assessment using nontraditional risk factors on patient health outcomes. The best available indirect evidence is mainly limited to studies evaluating the incremental value on discrimination and risk reclassification when adding ABI, hsCRP, or CAC to the FRS. We have much less evidence on the addition of these nontraditional risk factors to the PCE (compared to the FRS) and much less evidence to inform how these nontraditional risk factors improve calibration of traditional cardiovascular risk assessment. Therefore, the value of nontraditional risk factors to correct the over- or under-prediction of traditional risk assessment goes unanswered. Overall, ABI may improve discrimination and reclassification in women when the base model performs poorly, and CAC can moderately improve discrimination and reclassification with an unclear effect on downstream health care utilization. One large RCT shows that high-intensity statin therapy in individuals with elevated hsCRP and normal lipid levels can reduce CVD morbidity and mortality, but it is unclear if these benefits would not also be applicable to individuals with normal hsCRP. Treatment guided by nontraditional risk factors has not been evaluated against treatment guided by traditional multivariate cardiovascular risk assessment.