Comment and reply on: The cost-effectiveness of palivizumab for respiratory syncytial virus prophylaxis in premature infants with a gestational age of 32–35 weeks: a Canadian-based analysis

Summary

Correspondence on: The cost-effectiveness of palivizumab for respiratory syncytial virus prophylaxis in premature infants with a gestational age of 32–35 weeks: a Canadian-based analysis [Lanctôt KL et al., Curr Med Res Opin 2009; 24: 3223-37; doi:10.1185/03007990802484234]

Dear Sir,

In the article published in a recent issue of the journal by Lanctot et al.1, a cohort of 33 to 35 week gestational age infants was used to determine the expected benefit on mortality from use of palivizumab. The all-cause mortality rate over a 2-year period was 8.1% in children hospitalized with respiratory syncytial virus (RSV) infection versus 1.6% in children who were not, and these figures were used to calculate the expected reduction in mortality if RSV hospitalizations could be prevented. The vast majority of the deaths in the RSV cohort (76%) were due to sudden death of unknown cause2.

One conclusion would be that “fragile” infants are more prone to both RSV hospitalization and to sudden death of various etiologies, but it remains unknown to what degree preventing RSV hospitalizations will prevent these subsequent deaths.

In the sensitivity analysis, the authors also look at a much lower RSV-related mortality rate of 1.0% which was the rate in the original palivizumab trial3. However, 83% of the infants in this trial were born prior to 33 weeks gestation3 and would be expected to have a higher RSV mortality rate than would 33 to 35 week gestation infants. The RSV-related mortality rate was only 0.6% in the control group in a study of palivizumab in infants with hemodynamically significant congenital heart disease -- a group that most experts would consider to be at much higher risk of mortality from RSV than would be 33 to 35 week gestation infants4.

The main analysis assumes that prevention of RSV hospitalization will prevent asthma. There is no question children with asthma have a high incidence of RSV hospitalization, and it seems likely that prevention of these hospitalizations may impact the subsequent course of asthma, but it seems unlikely the entire condition can be averted by palivizumab. Again, the authors excluded any benefit on asthma in a sensitivity analysis and concluded palivizumab would still be cost-effective in infants at moderate risk for RSV hospitalization. However, had they performed a multiway sensitivity analysis assuming less benefit for prevention of asthma combined with a more realistic estimate for RSV-related mortality, palivizumab would no longer be cost-effective for other than the most high risk infants born at 33 to 35 weeks gestation.

Authors’ reply

Dear Sir,

Dr. Robinson raised several issues regarding our cost-effectiveness analysis of palivizumab in premature infants with a gestational age of 32–35 weeks in a Canadian setting1. We agree that there was a higher incidence of sudden or unexplained death in the respiratory syncytial virus (RSV) cohort as indicated in the Sampalis study2 than that found in both the IMpact and cardiac, short-term, randomized controlled trials3,4. Recognizing this, we performed a scenario analysis specifically to evaluate the impact of mortality rate assumptions on the incremental cost-effectiveness ratio (ICER).

Robinson stated that infants with hemodynamically significant congenital heart disease (CHD) are most likely to have a higher risk of mortality than the 32-35 weeks’ GA infants and quoted a RSV-related mortality rate of 0.6% based on the control group from the Feltes et al. CHD trial3. The overall mortality rate in the placebo arm of that trial was 4.2%. All-cause mortality, rather than disease specific, has been specifically recommended for use in cost-effectiveness analyses when disease specific rates are low5. Importantly, that study was not powered to demonstrate a statistically significant difference in RSV-related mortality rates, and not optimally designed to evaluate mortality as a primary outcome: patients not expected to survive for 6 months were specifically excluded and those enrolled were followed for less than 6 months. Therefore it is not surprising that the incidence was one of the lowest published.

While Robinson legitimately indicates that infants with hemodynamically significant CHD are at greater risk of mortality from RSV infection, studies from the pre-palivizumab era found mortality rates ranging from 2.5–37%6^-8. Moreover, based on a meta-analysis, Nuijten et al.9 in a cost-effectiveness study determined that the mortality rate for CHD infants hospitalized with RSV ranged between 4.0–4.5%. This is not only similar to the Feltes study but is indeed both more realistic and much higher than the RSV-mortality rate for infants of 32–35 weeks’ GA used in our sensitivity analysis. Therefore, we feel quite confident that the use of 1.0% in our sensitivity analysis is appropriate as a lower limit of mortality given the cohort of infants.

Robinson also noted that 83% of the infants in the IMpact trial were born prior to 33 weeks GA and thus may be more at risk of mortality than those born between 33 and 35 weeks GA4. The literature on late preterm infants actually demonstrates significantly higher early, late and post-neonatal mortality rates compared to term infants up to 1 year of life. (3 to 6 times higher)10,11. Perhaps this “fragile” cohort are therefore more susceptible to RSV illness than their preterm counterparts less than 33 weeks GA because of their immature immunologic system and incomplete pulmonary development that results in higher morbidity and mortality2,12. Taking this into consideration, the assumption that 1% is a representative lower limit for mortality for infants 32–35 weeks GA is appropriate and is most likely a conservative assumption for the study's cohort of interest.

We agree with Robinson that RSV prophylaxis may only be part of the answer to the complex asthma story. There is a paucity of information regarding the link between RSV and asthma in this particular sub-group and controversy on how asthma is best defined. While there is ongoing debate, it is hard to completely ignore the clinical evidence regarding the persistence of recurrent wheezing, chronic airway morbidity and impaired health-related quality of life from early childhood through 13 years of age following RSV lower respiratory tract infections in infancy13–17. Recognizing these issues, we presented the main results with and without the asthma assumption. If we reduce the benefits of palivizumab on asthma as Robinson suggested and use the mortality rate of 4.2% based on the Feltes trial, the ICER is approximately CAN$ $38,078 (2007) indicating that palivizumab is cost-effective and has moderate evidence for adoption. While that particular analysis was not in our paper, the probabilistic sensitivity analysis did vary mortality rate and asthma resource utilization (Table 11), and provides a range of values possible under different assumptions. Similarly, excluding asthma and applying the same mortality rate to the moderate and high risk groups in the Risk Scoring Tool, the ICER/QALY ranges from $64,880–$22,868 respectively providing moderate evidence for adoption. The results of both these analyses are in line with the sensitivity analyses in the paper and demonstrate the stability of the model.

Cost-effectiveness analyses always have an element of uncertainty, yet decisions have to be made based on available evidence. We used the Sampalis mortality rate of 8.1% for our base-case scenario because it is the only study that examined the outcomes of Canadian children who were hospitalized with proven or probable RSV over a duration of two years. Besides our base-case scenario, we also selected the most contentious issues to analyze. Using this approach, we quantified the possible effects of this uncertainty and, hopefully, highlighted important issues for future research. We thank Dr. Robinson for her comments regarding the need for more scientific data pertaining to this extremely vulnerable population.

Declaration of funding: none; Declaration of financial/other relationships: J.L.R. discloses that she has been an investigator in three studies of palivizumab that were funded by Medimmune and that she is listed as a coinvestigator on a grant related to palivizumab that will eventually be funded by Abbott, Canada. Acknowledgment: none.

Declaration of funding: Abbott Laboratories, Ltd. provided financial support for the analysis published in Ref 1; Declaration of financial/other relationships: The authors maintained control over the publication and the right to publish. Acknowledgment: No editorial assistance was provided during the preparation of this manuscript.