Dear Sir,
Re: Recommendation for the use of palivizumab as prophylaxis against respiratory syncytial virus in infants with congenital cardiac disease.
The above article was published in your journal of October 2003.1 We wish to express our disagreement with the recommendations. In our view, the evidence on which the recommendations are based is very weak indeed. Our disagreement is on a number of counts.
The authors recommend the use of palivizumab for the prophylaxis of respiratory syncytial virus infection in children with congenital heart disease. Unfortunately, the study on which this recommendation was based2 does not provide any supporting data. Indeed, the study does not mention the rates of infection in either the treatment or control arms. Hence, it is impossible to determine if treatment has any effect on prophylaxis. Instead, as indicated in the title of the paper, it only claims to reduce hospitalisation following infection by the respiratory syncytial virus in such children.
The evidence on which the recommendation for treatment of children with heart disease is based relates to only one statistically significant finding, namely, a relative reduction of risk in hospital admissions. They reported a “p” value indicating significance without looking at absolute figures concerning reduction of risk, or numbers needed to treat to save one episode. As our Table 1 shows, relative reduction of risk can sometimes appear misleadingly significant when looked at in isolation.
Table 1.
Four hypothetical studies each with an RRR of 50%
Even if one accepts the validity of relative reduction of risk for the end point, as they have done, it was still necessary to treat 22 children before one could be prevented from being admitted to hospital. Considering that at least 5 injections in each winter season are required, at a cost of £829.00 per injection, the authors are recommending that we spend over £90,000.00 to prevent one baby from being admitted to hospital. In the absence of any data suggesting any benefit after treatment in terms of death or serious morbidity, we believe that this is a questionable use of the limited resources available within the National Health Service. More robust evidence is therefore required before recommendations can be made for routine clinical use of this very expensive drug. This evidence should include a detailed assessment of the financial aspects of this treatment in the United Kingdom, particularly the ratio of cost to benefit.
In Table 2, we include our estimates, based on the data in the paper, of the cost of reduced admission to the intensive care unit, and decreased need for mechanical ventilation, which are the other secondary end points in the study. The figures for these are of even more concern, although the reduction of risk for these did not achieve significance.
Table 2.
Analysis of primary and secondary end points in the study.
Finally, the declaration of interest stated that a number of the authors were also investigators in the study. There is nothing in the declaration of interest to suggest that the authors benefited from the study, and we have no doubt of their integrity. Probity might have been better served, however, had the advisory group been completely independent. Perhaps the British Paediatric Cardiology Association should exercise more caution in the future in situations like this, where potential conflict of interest might exist.
