In this March issue of BJSM, Wilson and colleagues investigate sudden cardiac death:
This emotional and very important area of sports medicine always raises the issue of how many children how have cardiac abnormalities may need to be disqualified from sport to save one life.
Editorial Board member Jon Drezner posted the following comments relating to this paper (originally posted to the AMSSM listserve):
The studies on ECG screening are mounting fast with improved and more specific ECG criteria that lower the total positive (and thus false positive) rate. Pelliccia et al (Euro Heart J 2007) recently described their experience reviewing 32,652 screening ECGs primarily in young amateur athletes (median age 17). Distinct ECG abnormalities suggesting cardiac disease were present in only 4.8%. This lower total positive rate was after acknowledging that a prolonged PR, incomplete RBBB, and early repolarization patterns are essentially normal/common findings in athletes and not indicative of cardiac disease. Other refinements to the definition of “abnormal” for a screening ECG are also being recognized. A recent study presented at the 2007 AHA Scientific Sessions by Melacini et al (Marginal overlap between ECG abnormalities in patients with HCM and trained athletes: implications for preparticipation screening) found that voltage criteria alone for LVH (without ST depression, T wave inversion, or pathologic Q waves) is a common finding in trained athletes and unlikely to be indicative of cardiac disease.
Two recent studies have used modifications of the Corrado criteria and found a substantially lower (about 2%) total positive rate. In a recent study by Wilson et al. (BJSM 2007) out of the U.K., 2,720 national/international athletes and physically active school children (mean age 16) were screened using personal & family history, exam, and ECG. They found a total ECG positive rate of only 1.5%. Nine athletes (0.3%) were identified with cardiac disease known to cause SCD (WPW, LQT1, ARVC, RVOT), and none of these 9 cases were symptomatic or would have been identified by personal or famhx. (see abstract below). I found it interesting that the true positive rate found was consistent with the AHA estimate regarding the prevalence of cardiac disease known to cause SCD in young athletes ( 0.3% or 3 in 1,000).
Also recently, Joseph Marek from the Midwest Heart Foundation presented at the 2007 AHA Scientific Sessions their findings of screening over 12,500 high school aged individuals with ECG. They also used modified Corrado criteria for defining abnormal ECGs. Their total positive rate was only 2%. This is the largest study in the U.S. and the first I am aware of to apply the Italian criteria to a U.S. population.
Another way to look at the numbers is to model a screening program for 50,000 high school freshman athletes: (1) Incidence of SCD is approx 1:50,000 (based on current data from the U.S. Sudden Death in Young Athletes Registry by Barry Maron); (2) Prevalence for any cardiac disease known to cause SCD in young athletes is estimated to be 0.3% or 3 in 1,000 (from the 2007 AHA Scientific Statement on preparticipation screening); (3) A screening ECG will suggest about 60% of silent CV dz known to cause SCD.
Assume 2-5% total positive rate (based on studies above using updated ECG criteria) = 1,000 to 2,500 total positives. 150 potential true positives (prevalence 0.3%), but only 100 (about 60%) true positives with CV dz suggested by ECG (0.2%). That leaves 900 to 2,400 false positives (1.8% to 4.8%).
In other words, we would need to temporarily disqualify/work-up 9 to 24 kids with false positives to identify 1 kid at risk for SCA, or we would need to disqualify 99 kids with CV diagnosis to potentially prevent 1 death (in the first year of screening). However, each year it is possible that an additional SCD event will be prevented through disqualification (this is shown in Corrado’s study from Italy). Assuming high school participation for 4 years (200,000 person years of athletic participation in this model), we might expect 4 cases of SCD (incidence 1:50,000 per year). Since ECG did not capture all of the silent CV dz (only about 60%), then screening 50,000 high school freshmen with ECG would likely prevent 2-3 deaths (of the predicted 4) through disqualification of 100 kids with identified cardiac dz. In other words, we would need to disqualify 30-50 kids with identified CV dz from high school athletics to prevent 1 death.
I recognize these calculations are rough, but I’m trying to pull from the most recent studies and apply to our setting.
Anyway, new studies, more accurate and rigorous ECG criteria, and lower total positive rates. Over time with confirmatory studies and a better understanding of disease prevalence, this will change our calculations on false positives, disqualifications, cost, and lives saved, and quite possibly compel us to revisit our recommendations on the role of ECG in the screening process.