Dear Sir,

We thank Dr Onuzo et al. for their forthright criticism of our recommendations for the use of palivizumab in infants with congenital heart disease, and you, Sir, for the opportunity to respond.

The main thrust of the criticism appears to be that we have not considered the cost-effectiveness of this treatment, though we clearly state in our article that evaluation of cost–benefit is still required.

It cannot be denied that infection of vulnerable infants with congenital cardiac disease by the respiratory syncytial virus represents a threat to their well-being and survival. Paediatric intensive care units around the country prepare themselves each year for the influx during the autumn and winter months of children infected in this way, and every paediatric cardiology unit has experience of children undergoing prolonged intensive care, with some patients dying as a result of this infection.

There are a number of points raised in their letter that we will deal with in order.

• Although they express disagreement with the recommendations they do not clearly state the aspects with which they disagree.
• We refute the assertion that the recommendations are based on weak evidence. The study of Feltes et al.² was a double-blind, randomised, placebo- controlled trial. As such, it represents the highest level of evidence, short of a meta-analysis of several randomised-controlled trials. This provides us with evidence of both the safety and efficacy of this treatment.
• They say that we recommend the use of prophylaxis in children with congenital cardiac disease, whereas our recommendations largely concern infants and then only those with haemodynamically significant congenital heart disease, pulmonary hypertension and children with cardiomyopathy receiving treatment. We also state that consideration should be given to infants requiring admission for medical or surgical intervention during the season of infection with the virus, and children over the age of 1 with complex cardiac disease. In fact, it is clear that our recommendations will help the clinician justify more limited use of this expensive treatment. Our group were at pains to define patients who would not be regarded as likely to benefit from its use, as much as to identify those who likely would.
• We disagree that the study of Feltes et al.² does not provide data to support these recommendations. The study demonstrated that palivizumab recipients had a 45% relative risk reduction in hospitalisations for those infected by the virus. Though this does not tell us the rates of infection in the whole population studied, it strongly suggests that palivizumab has a positive impact on the burden of the disease in this group of vulnerable patients. The study does show reduction in stay in the intensive care unit, assisted ventilation, and death. To demonstrate significant reduction in these areas specifically due to infection with the virus, however, it would be necessary to perform a study on over 10,000 children with congenital cardiac disease so as to achieve an 80% power at the 5% level. For this reason, hospitalisation because of infection by the virus was the primary end point of the study.
• We note the criticism regarding the use of relative risk reduction as a statistical tool. We agree that it can be misleading when looked at in isolation. The incidence of hospitalisation because of infection by the virus in this group of 1287 patients, however, is clearly stated as involving 5.3% of the 639 patients who were treated, as opposed to 9.7% of the 648 who received placebo. The relative reduction in risk, and 95% confidence intervals, is clearly stated, along with the p value obtained using Fisher's exact test. It is self-evident that, in order to reduce the incidence from 10% to 5% (approximately), it is necessary to treat 20 patients so as to prevent one hospitalisation. These analyses undoubtedly should be part of the cost–benefit analysis that we stated were needed. Assessment of the number needed to treat analyses of the pharmacoeconomic issues from the payer's perspective, rather than society's. They do not include the costs of long-term consequences to an infant and their parents, including death from infection by the virus, or death or morbidity due to delayed correction or palliation of a cardiac lesion. As clinicians, we must be mindful of the wider implications of infection by the respiratory syncytial virus in our children, including mortality and morbidity. In this study, 6 of the 97 patients admitted to hospital died (6.7%), which illustrates the seriousness of infection by the virus in this group of patients.
• We believe that the costs as quoted by Dr Onuzo et al. are excessive compared to figures we have been quoted from our pharmacies. In addition, there are simple savings to be made by batching “at-risk” patients, so maximising use of each ampoule.
• Whilst we agree that we have responsibilites for judicious use of the resources available within the National Health Service, we refute the suggestion that we are recommending routine use of this expensive treatment. Our objective was to provide guidance to paediatric cardiologists and paediatricians faced with decisions regarding appropriateness of passive immunisation of individual patients under their care. It should be noted that palivizumab is being widely used in general paediatric practice for premature infants, infants with chronic lung disease, and those with immunodeficiency. It is approved for use by the Federal Drug Administration in the United States of America, and also for use in the European Community, both for these patients and for children with congenital cardiac disease.
• We are disappointed that Onuzo and his co-signatories have chosen to question our individual probity, as well as that of the British Paediatric Cardiac Association. The members of the working party were chosen because of their interest and expertise in the subject, and this inevitably included paediatricians and cardiologists who had contributed to that part of the study undertaken in the United Kingdom. Views were also sought from clinicians unable to attend the working group, as indicated in the acknowledgements. The authors of the letter are surely not suggesting that any participant in a commercially sponsored, randomised-controlled trial should be denied the opportunity to contribute to future recommendations in the area of their work and expertise?

Finally, we would suggest that professional bodies, including the British Paediatric Cardiac Association, have a duty to take an active role in the evaluation of new treatments that may benefit patients.