European Heart Journal Advance Access originally published online on December 8, 2006
European Heart Journal 2007 28(1):137-139; doi:10.1093/eurheartj/ehl415
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Cost effectiveness of neonatal ECG screening for the long QT syndrome: reply
Department of Cardiology
University of Pavia
IRCCS Fondazione Policlinico San Matteo
Pavia, Italy
E-mail address: pjqt{at}compuserve.com
Department of Computer Science and Systems
University of Pavia
Pavia, Italy
We are grateful to Dr Van Hare and his well-respected colleagues for raising important questions related to our article.1 In writing this article, we focused on the cost-effectiveness issues and we took for granted some important considerations very familiar to us but perhaps less so to others. Their letter offers a unique opportunity to clarify these significant points.
The skepticism on the issue of costs is unjustified. Costs were considered from the point of view of the Italian National Health System (NHS); all costs listed are those established within the NHS and used for hospital reimbursement. To cover the possible variations within European countries, we conservatively varied every single parameter by ± 30%. The fact that public health-related costs are lower in Europe than in the USA is no surprise. As to the no costs associated with training adult cardiologists to read neonatal ECGs defined as ‘unrealistic’, we stand by our statement. In Italy, as throughout the Europe, hospital-based cardiologists are paid by the State and operate within the NHS. When they become ‘cardiologists’ they are supposed to have learnt to read ECGs, independently of the age of the patient. If they have forgotten (or never properly learnt) how to read a neonatal or paediatric ECG, this is their problem. They should go back to their books and study. The State will not pay for their additional instruction (taxes to attend medical schools in Italy are negligible and formal training to become a certified specialist in cardiology is paid by the State, for those admitted). Furthermore, the European Society of Cardiology has published specific Guidelines exactly for this purpose.2 If adult cardiologists will follow them, they will identify most problems. Additional voluntary training will be useful and possible, at no cost.
It is true that no single test exists to rule out LQTS in general and specifically in an asymptomatic individual with an initially prolonged QT interval, but diagnostic defeatism is unwarranted. We, as cardiologists, have been able to diagnose and exclude (with a reasonable degree of medical certainty) LQTS well before the advent of molecular cardiology. If an asymptomatic child (say, 3 years old) without a family history of LQTS and with a normal QT interval, confirmed in repeated ECGs during the last couple of years, comes to us and the mother says that in the first month of life the infant had a QT prolongation (subsequently vanished and never observed again), who among us would diagnose LQTS and recommend beta-blockers? No one, if competent. Thus, the risk of an occasional and transient QT prolongation in the first month of life leading to a ‘perpetual’ diagnosis of LQTS with life-long medical treatment (and with the attendant catastrophic scenarios imagined by Van Hare et al.) is almost non-existent if the cardiologists use knowledge and common sense.
Van Hare et al. rightly call attention to the issue of false positives, but correct quantifications are essential. They quote our figures but make a gross error as they refer to ‘a positive rate of 1% for a QTc cut-off of 470 ms’. The findings we reported,1 based on 33 000 ECGs analysed prospectively,3 were 0.9/1000 (almost 1/1000). Actually, the now complete analysis on the 45 000 infants enrolled shows that the prevalence of infants with a QTc >470 ms has further decreased to 0.7/1000. Thus, the magnitude of the problem of infants with markedly ‘prolonged QT interval’ is quite different from the one portrayed by Van Hare et al.
In our analysis the rate of 1% has been assumed for the false positives at the first ECG (QTc >440 ms). Probably, the label ‘Markov model for inappropriate therapy’ in Figure 1, where the decision tree is represented, was not fully appropriate and may have generated misinterpretations for which we apologize. In fact, the entire model considers the continuation of the diagnostic procedures. The 1% false-positive rate at first ECG decreases to 0.03% (3/10 000) when considering the entire diagnostic process (two or more ECGs, Holter, family history, and so on), as recommended.2 Owing to space limitation, we could not include a clarifying figure depicting the final phase of the Markov process related to the false-positive cases. Our model obviously takes into account the cost of every additional test done for false-positive individuals and for the treatment of the 0.03% among them who will start beta-blockers. Moreover, the Markov process considers that 80% of the false-positive children under treatment will stop the therapy within few years.
Using our correct figures for prevalence (1/2500), sensitivity (80%), and false-positive rate (0.03%) for the entire diagnostic process, the positive predictive value calculated applying Bayes' rule4 is 51.7%. This is quite at variance with the positive predictive value ‘of only 3%’ incorrectly calculated by Van Hare et al. (in their calculation, there must have been an additional error besides the 1% vs. 1/1000). If one considers the overall financial national burden, in Italy the entire screening on 550 000 newborns would cost 
23 million Euros/year.
Van Hare et al. are expert paediatric cardiologists. They would not take a QTc of 450 ms in a 3-week-old infant as a sign of LQTS and initiate beta-blocker therapy for life. Neither would we. The European Guidelines,2 to which two leading American paediatric cardiologists (Dr A. Garson, Jr and Dr V.L. Vetter) have significantly contributed, have been careful in outlining the sequence of steps to be followed once the first ECG has revealed a QTc exceeding 440 ms. We truly recommend that Van Hare et al., as anyone interested in this issue, read them carefully. Some points are critical. If the second ECG is normal, the case is dismissed; if it shows a QTc between 440 and 470 ms, then several specified checks, including family history, are suggested and if negative, ‘no treatment is currently recommended’. We did specifically mention that ‘even in infants with a very prolonged QTc in the first month of life, the ECG may normalize’ and that ‘if subsequent ECGs and diagnostic procedures do not confirm the presence of LQTS, it is logical to progressively withdraw therapy’.2 The bottom line is that a careful adherence to the Guidelines will lead to a very small number of infants treated with beta-blockers and most of them will truly need therapy.
Their final point addresses ‘the psychological and emotional impact on the families incorrectly labelled as having LQTS’. First, this applies to only three infants per 10 000 and for the 2–3 years necessary to rule out LQTS (as mentioned earlier). For those with a QTc between 440 and 470 ms at the first ECG, using the procedure mentioned earlier and a cautious approach, anxiety can be reduced to a minimum. In any case, however, one has to look at the counterpart. We have seen, over the last 30–35 years, the eyes of too many parents who lost their infants because of undiagnosed LQTS to be swayed by ‘some transient anxiety’. Any choice carries its errors, but there is no remedy for sudden death.
‘Mass screening’ looks like mandatory screening. We never thought of that. We just recommend that our countries make available a service of neonatal ECG screening for the interested families. What we consider, however, a precise responsibility of neonatologists and physicians is to inform the families of a newborn about the existence of LQTS, its prevalence, its dangers even in the first year of life, its effective treatments, and the fact that it can be diagnosed or suspected by a simple ECG. At that point, it will be the free choice of individual families to have an ECG made or not. In some countries, a free screening may be offered, in others not. What matters to us is that the parents are provided with the essential information to make a reasoned choice. We believe that this approach will contribute to reduce the number of avoidable deaths due to LQTS in infancy or later on in life.
Van Hare et al. ‘cannot support a recommendation for mass screening’, a legitimate position. If this position, however, was based on the grossly incorrect figures and views reported in their letter (1% of infants with QTc >470 ms, a large number of false positives diagnosed with LQTS and treated for life with beta-blockers), perhaps our reply may help them to reconsider. When confronted with the correct facts, we trust that they will recognize that proper information on LQTS should be given to the parents of a new baby and that the consideration of a service, available to all families, to provide an ECG during the first month of life—at least for those countries endowed with an NHS—is not without foundation.
References
- Quaglini S, Rognoni C, Spazzolini C, Priori SG, Mannarino S, Schwartz PJ. (2006) Cost-effectiveness of neonatal ECG screening for the long QT syndrome. Eur Heart J 27:1824–1832.
[Abstract/Free Full Text] - Schwartz PJ, Garson A Jr, Paul T, Stramba-Badiale M, Vetter VL, Villain E, Wren C. (2002) Guidelines for the interpretation of the neonatal electrocardiogram. Eur Heart J 23:1329–1344.
[Free Full Text] - Goulene K, Stramba-Badiale M, Crotti L, Priori SG, Salice P, Mannarino S, Rosati E, Schwartz PJ. (2005) Neonatal electrocardiographic screening of genetic arrhythmogenic disorders and congenital cardiovascular diseases: prospective data from 31 000 infants. (Abstract). Eur Heart J 26:Suppl., 214.
- Weinstein MC and HV Feinberg. (1980) Clinical Decision Analysis(WB Saunders Company, Philadelphia, PA).
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J N Johnson and M J Ackerman QTc: how long is too long? Br. J. Sports Med., September 1, 2009; 43(9): 657 - 662. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||