The implications of a relationship between prolonged QT interval and the sudden infant death syndrome.

Abstract

We read with interest the article on the proposed association of prolonged QT interval and the sudden infant death syndrome (SIDS) by Schwartz et al.1 Recognizing the previous failures to determine cause(s) for SIDS and the risks of changing practice in response to one study, we wish to take this opportunity to place the findings of the study in perspective and reflect on directions that might be pursued.Although the mechanism for SIDS (the most common cause of death from 1 month to 1 year of age) remains unknown, certain factors, such as smoking during pregnancy and sleep position of the infant lter the risk of SIDS.2–7 The hypothesis that there is a relationship between SIDS and prolonged QT has been previously raised by Schwartz8 and other authors,9–12 and disputed by some,13–15 but the recent data by Schwartz et al1 are the most compelling to date. Schwartz et al obtained electrocardiogram (ECG) recordings 3 to 4 days after birth in 34 443 newborn infants, 24 of whom died within the first year after birth with a diagnosis of SIDS. The authors found that these 24 infants had an average QT interval (corrected for heart rate) that was longer than in 9725 randomly selected living infants and longer than in 10 infants who died of other causes. Accordingly, this report raises several important questions.What is the cause of the long QT interval in the infants?Schwartz et al1 suggest that the prolonged QT might be caused by: a) different rates of development of the right and left sympathetic nerves rendering the infant susceptible to an arrhythmia during sudden increases in sympathetic activity8,16,17; or b) mutation in one of the genes encoding sodium or potassium channels, as has been identified in some families with a high incidence of syncope and sudden death. Prolonged QT interval can also be caused by myocarditis, hypocalcemia, or some medications, (eg, cisapride or quinidine), although none of these has been associated with SIDS. In any of the causes of long QT interval the patients characteristically develop torsade de pointes, a specific form of ventricular tachycardia.Is there additional support for an association between prolonged QT interval and SIDS? Another report that lends support to this association is by Maron et al,9 who studied 42 sets of parents who had an infant with SIDS, and found that in 11 of the sets, at least 1 parent had prolonged QT. Furthermore, in 8 of the families with 1 parent who had prolonged QT, 9 of the 23 SIDS siblings also had prolonged QT interval, consistent with an autosomal dominant pattern of inheritance. However, the Maron report does not have measurements of QT in the SIDS victims. The article by Schwartz et al does not have QT measurements at or near the time of death, but their earlier study shows that QT intervals increase in early infancy,18potentially rendering the neonates with long QT even more susceptible to an arrhythmia at 2 to 3 months. Sadeh11 also studied ECG recordings from 10 infants who subsequently died of SIDS from a large cohort of infants studied in England. They found that 5 of 10 SIDS victims failed to shorten QT interval appropriately with increasing heart rate, which would also leave them susceptible to an arrhythmia. In contrast, Southall13 studied 7254 neonates, 15 of whom died from SIDS, and none of the victims had a long QT interval.How compelling is the evidence that SIDS is caused by a cardiac arrhythmia? Currently there is no other evidence for this, and some evidence against. a) A few SIDS children have had heart rate monitors attached at the time of death, and all showed marked bradycardia (possibly from hypoxia), but none had ventricular tachycardia.19 b) The autopsy examinations of infants with SIDS are most consistent with chronic or repetitive hypoxia.20 c) Autopsy also shows subpleural hemorrhages, which are most compatible with airflow obstruction.20These observations are more compatible with respiratory rather than cardiac dysfunction preceding death. Moreover, one might expect that, at least by serendipity, torsade de pointes would be observed in some infants in hospital if prolonged QT were a common mechanism contributing to SIDS, yet we can find no reports of this. Although it is conceivable that apnea could be treated and not reported, torsade de pointes is more likely to be noticed. Thus, even if there is an association between SIDS and long QT, it would be very premature and potentially erroneous to change care practice before there is documentation that an arrhythmia is a mechanism for SIDS.Are there other explanations for an association between prolonged QT interval and SIDS? a) If the long QT intervals are not caused by genetic defects or drugs, it remains possible, although not well-supported by data, that there is an imbalance between the development of innervation by the right and left sympathetic nerves. b) It is also possible that prolonged QT interval is an independent marker of pathology, rather than the proximate cause of death in infants with SIDS. The brainstem of infants with SIDS shows abnormalities of certain medullary nuclei associated with respiratory control.21,22However, the abnormal receptor binding in the medulla in SIDS victims identified by Filiano and Kinney,21 affects circulatory as well as respiratory control, and it is quite possible that abnormalities of other central nervous function could simultaneously disturb both systems. c) Alternatively, conditional factors eg, sleep state or body position, might alter QT interval and susceptibility to SIDS in a dependent fashion. Quiet sleep prolongs QT interval23 and it is possible that the change in posture could affect the proportion of time in quiet sleep and thereby alter the QT interval. ECG recordings are customarily obtained while the infant is supine, possibly altering the likelihood of detecting prolonged QT; thus, we might speculate that the recently observed reduction in SIDS with supine posture could relate to a decrease in QT duration.What is the utility of large scale screening of ECG to detect infants at risk for SIDS? The data provided by Schwartz et al suggest that measurement of QT would be a poor means for identifying infants with SIDS. There were 34 418 infants with no SIDS in their study; 2.5% of these, or 860, had corrected QT interval >440 msec, their cut-off value for long QT. These 860 infants would befalse-positives in any screening process. Of the infants with SIDS, 12 had a normal corrected QT, and would befalse-negatives (if we consider SIDS a unitary disease); the other 12 infants with SIDS had a prolonged QT and would betrue-positives. These estimates give, at best, a positive predictive accuracy (true-positive/true-positive + false-positive) of [12/(12+860)] × 100 = 1.4%. Therefore, mass screening would identify ∼99 infants with a good outlook for every 1 who might have SIDS, and this is a poor ratio for any screening test.Where do these observations leave us? We propose the following: