番茄社区

Widespread screening of asymptomatic children for inherited diseases that cause sudden cardiac death may not be ready for prime time, according to a meta-analysis.

Overall, the prevalence of ECG- and/or echocardiography-detected hypertrophic cardiomyopathy and long QT syndrome among study subjects was low, Laurel K. Leslie, MD, MPH, from Tufts University School of Medicine in Boston, and colleagues found.

In addition, while the negative predictive value of ECG to rule out hypertrophic cardiomyopathy and long QT syndrome was nearly 100%, the positive predictive value varied depending on cutoff values for sensitivity and specificity, they reported online in Pediatrics.

Because of the low prevalence and the variation in false-positive rate, "further cost- or comparative-effectiveness analyses will be necessary to determine whether screening programs to detect sudden cardiac death in asymptomatic children should be promoted as public health policy," researchers wrote in conclusion.

The inherited arrhythmias examined in this meta-analysis are the two most common disorders in children that are detectable by ECG. (Researchers also examined ECG screening for Wolff-Parkinson-White syndrome, but because of study limitations they assumed the sensitivity and specificity were one and the positive and negative predictive values were perfect and did not discuss it further.)

For a screening program to be acceptable, it should be "highly discriminatory," the authors noted. It should identify the disorder in all affected patients (sensitivity) and rule them out in healthy children (specificity).

But "low prevalence and imperfect sensitivity and specificity estimates could result in inefficient screening strategies with unanticipated societal and economic costs," researchers said.

To help define boundaries of the theoretical utility of ECG and/or echo as screening tests for these disorders, the investigators included 30 studies in their analysis. They noted the substantial variation between the studies.

The summary estimates of the prevalence of the disorders per 100,000 asymptomatic children were:

• Hypertrophic cardiomyopathy -- 45 (95% CI 10 to 79)
• Long QT syndrome -- 7 (95% CI 0 to 14)
• Wolff-Parkinson-White syndrome -- 136 (95% CI 55 to 218)

When the sensitivity and specificity cutoff values were equally weighted with the prevalence of the disorders, the negative predictive value for ECG, echo, or ECG/echo to rule out cardiomyopathy or long QT syndrome was nearly 100%.

But the positive predictive value was less than 1%. In addition, the number needed to screen to detect one case ranged from 2,611 (using echo to detect hypertrophic cardiomyopathy) to 16,592 (using ECG to detect long QT syndrome).

Consequently, the number of false-positives estimated when detecting one case ranged from 389 to 2,323, respectively, and the number of false-negatives per 100,000 screened ranged from seven to one, respectively. The use of echo alone, rather than in combination with ECG, or ECG alone, yielded slightly better values for detecting hypertrophic cardiomyopathy.

When Leslie and colleagues gave more weight to specificity (maximum specificity 99%), the positive predictive power, although still low, ranged from 2% to 21%, while the negative predictive value fell imperceptibly from 0.9999 to 0.9996 (ECG for hypertrophic cardiomyopathy) and 0.9998 (echo and/or echo/ECG for hypertrophic cardiomyopathy).

When specificity was given more weight, the number needed to screen to detect one case rose dramatically as the false-positives decreased and the false-negatives increased.

For example, an increase of 21% in the positive predictive value of using ECG to screen for long QT syndrome resulted in the number needed to screen to detect one case rising from 16,592 to 134,771. In addition, the sensitivity was reduced, which resulted in missing more diseased individuals.

Interestingly, the area under the receiver operating characteristic curves (AUC) for ECG were 0.91 for detecting cardiomyopathy and 0.92 for long QT syndrome, and 0.89 and 0.88 for echo and echo/ECG, respectively, to detect hypertrophic cardiomyopathy.

The AUC indicates that "ECG and/or echo are statistically acceptable screening tests," the authors said. However, the low prevalence makes screening on a population-wide basis fraught with concerns. These include high rates of false-positives, concomitant with additional downstream testing costs, potentially unnecessary therapies, and unwarranted child and family anxiety.

When the authors calculated how their findings would apply to population-based ECG screening, they found a positive predictive value of 41% when specificity was maximized, which was better than 1% but still left a "false-alarm" rate of 59%, they said.

Leslie and colleagues noted that their results did not take into account potential changes in mortality that could result from earlier diagnosis and treatment. Other limitations to the study include the potential for publication bias, the large variation between studies, and the lack of data on individual physicals and family history.

"Despite these limitations, this study provides an important starting point for evaluating sudden cardiac death screening programs," they concluded. However, they cautioned that more data are needed before adopting such programs as public policy.

From the American Heart Association:

Comment