Discussion
Our study found that CV testing among patients with CV risk factors and without known heart disease occurs in about a third of patients at the highest level of SCD risk. Those in the highest SCD risk categories had a trend for more testing but this peaked at approximately one third of the patients. This higher rate of testing in those at highest SCD risk was associated with a 7% all cause mortality rate and a 45% chance of developing sudden death over 10 years. The strengths of these findings are supported by the use of data from two large academic centers, the diversity of the sample, the large cohort and the consistency of the findings.
Screening populations for SCD is a significant challenge.1, 15 Out of hospital cardiac arrest is a leading cause of death among adults in the United States (approximately 300,000 events per year).1 Even though the risk of SCD is higher in patients with heart failure and/or advanced cardiomyopathies, the majority of SCD cases occur in the general population with cardiovascular risk factors and no known heart disease.23 Identifying people at risk of SCD without known heart disease is therefore challenging.15 Left ventricular ejection fraction (LVEF) is a strong predictor of SCD, but it is neither sensitive nor specific.17 Furthermore, population screening for ventricular function is not practical nor recommended for routine clinical practice.4, 16 This inverse relationship between risk (or incidence) and actual cases of SCD was initially highlighted by Myerburg et al.2 To date, there has been no systematic attempt to address SCD in the general population.
The first major breakthrough to enable risk stratification for SCD in the general population was the development of a simple clinically based risk score that could provide risk stratification for SCD in the general population. Bogle et al developed a risk model from ARIC and validated it in the Framingham study. Model covariates included age, sex, total cholesterol, lipid-lowering and hypertension medication use, blood pressure, smoking status, diabetes, and body mass index. The profile of risk score by decile yielded an exponential curve with the bulk of the SCD risk in the upper two deciles. While other risk scores have been proposed, they all include variables that may not be uniformly available in a primary care setting, such as an ECG. In our primary care cohorts, 90% of patients had all the variables required to calculate this risk score. One can argue that all of the variables in this risk score should be documented in the medical record.
In the report by Bogle et al, the predicted and observed 10-year rates of SCD in the upper two deciles were in the 2-6% range. While this is a substantial risk, it is not actionable as it is still fairly low and there are no identified targets for treatment. With this background, the current study was designed to identify how frequently these high risk patients are undergoing cardiac evaluation. These patients have a very high burden of cardiovascular risk factors. It is therefore possible that these patients have significant subclinical cardiovascular disease that can only be identified by cardiac testing. We demonstrate a very low rate of cardiac testing in this high risk population, and therefore identify a potentially practical opportunity to identify subclinical cardiac disease that could be targeted for treatment that may ultimately prevent SCD.
While this approach requires careful evaluation to establish its efficacy, this is a paradigm shifting approach. Currently, the three main bodies writing preventive cardiac guidelines do not support screening of asymptomatic individuals. Second, it is not cost-effective to screen for CV diseases in all patients seen. Third, insurance will not reimburse for asymptomatic testing and physicians do not order them because patients are not willing to pay out of pocket. However, echocardiograms in asymptomatic primary care patients have been able to identify incident cases of systolic and diastolic dysfunction which in turn lead to changes in treatment and this screening strategy has been proven to be cost-effective.5, 6 As it would be cost-prohibitive to perform testing on all patients seeking care in the United States, testing in the highest risk deciles may be more practical. There is a clear need to act on the diagnostic gap identified in this study.
The nature of the further cardiac testing to recommend in these high risk deciles is certainly not established. Our group has found that testing seniors in a primary care practice using echocardiography can identify new HF patients and those with heart failure with preserved ejection fraction.5, 6 We also have documented that by informing the clinicians of the echocardiogram findings there is medication intensification, improvement in CV risk factors and lower operative costs.18 Also adding echocardiographic information to the SCD score improves the ability of the combined score to predict mortality.19
There are several limitations that deserve mention. First, we relied on ICD codes to identify comorbidities and used this as an exclusion criteria, this could have excluded patients with HF or CAD when the condition was not present. However, we conducted a chart review in a small sample and found that the accuracy of the codes was 100%. Second, we used a cross-sectional design to evaluate the outcome and this could have undercounted the number of procedures performed. However, we used period prevalence to calculate the procedures over time. Third, patients could have had their procedures done at other institutions and this could have led to undercounting of the procedures. Fourth, the generalizability of our findings is limited to the two medical centers included in the study.
In conclusion, this study provides evidence that the majority of patients from the general population at elevated risk for SCD based on a novel SCD risk score do not undergo any CV testing. We propose a new paradigm of further cardiac screening for these asymptomatic patients to identify subclinical cardiac disease that leads to SCD. Future studies should evaluate what the appropriate testing is and the therapeutic strategies to implement for those at risk.
Table 1: Baseline characteristics by SCD risk: A) University of Miami B) Vanderbilt University
A)