Researchers use a comparative study to model RSV prevention in children

By Dr. Liji Thomas, MDNov 28 2022Reviewed by Aimee Molineux

Now that the coronavirus disease 2019 (COVID-19) pandemic has been around for almost three years, scientists are looking at other resurgent respiratory viruses that are taking a heavy toll on human health following the removal of mask-related and other public health protections against virus transmission. Among these viruses is the respiratory syncytial virus (RSV).

Using computational modeling, a new paper compares the benefit of maternal vaccines versus the administration of RSV-specific monoclonal antibody (mAb) therapy against the virus. They used five static and dynamic models to examine a hypothetical cohort of 100,000 babies to evaluate the outcomes following seasonal and year-round intervention programs. The input data came from the evidence collected from either vaccine or mAbs during the study period.

Introduction

RSV causes well over three million acute lower respiratory infections (ALRIs) in children, and another 3.6 million hospitalizations, among children of five years or less, as per 2019 figures. Palivizumab is the only licensed intervention and is a mAb given monthly by injection to high-risk infants to prevent the infection. However, its high cost limits its clinical use.

Many other interventions are being developed, such as nirsevimab, which is also a mAb but has long-term efficacy after a single dose. It is currently undergoing accelerated review. Other drugs, including another mAb and a couple of maternal vaccines, are in phase 3 trials. Their clinical usefulness will depend, of course, not only on their efficacy but also on their cost-effectiveness analyses (CEA), which inform many new immunization protocols.

Several models have been introduced to study the cost-effectiveness of preventive programs for RSV but are limited by restricted data and varying results. This study aimed to produce modeling estimates using the same data and assumptions from several models. The resulting outcomes could be compared with greater utility than those from differing studies.

In all five models used here, the RSV season was assumed to begin in October and end in April, as is typical of Europe. The population of interest comprised pregnant women and infants, while the measure of cost-effectiveness was the incremental cost-effectiveness ratio (ICER) in terms of quality-adjusted life years (QALYs).

The three static models estimated the number of cases that would have required a physician visit or hospitalization (“medically-attended”) averted by maternal vaccination or mAbs offered year-round or seasonally versus the number of cases that would have occurred without intervention. The year-round maternal vaccination program was estimated to cover two-thirds of eligible women and 94% of infants at birth.

The seasonal program covered 44% and 94% of mothers and infants, respectively, while the addition of a catch-up program achieved the same coverage of infants born less than six months before the RSV season start date.

What did the study show?

With the static models, approximately one thousand cases were estimated to be averted by the maternal vaccine, amounting to a direct saving of a million pounds in medical costs and a third more in non-medical costs. Simultaneously, babies less than a year old were estimated to have gained 4-5 discounted QALYs, for each year of intervention.

Coming to the use of mAbs, the direct and indirect savings came to four million and 1.5 million pounds, respectively, with a much higher gain of 20-25 QALYs.

Using dynamic modeling, the number of cases averted by maternal vaccination comes to about 400-700, fewer than the estimate arrived at by static modeling. With mAb intervention, the number of cases averted is about 3,300 to 4,600. The reductions are due to the assumption that protection afforded by either intervention wanes over time so that the number of cases averted is less within two months after birth.

The dynamic models could not reflect the hospitalizations before and after the seasonal peak. Interestingly, one of these models showed that the age when RSV cases peaked changed to affect children aged 2-5 years following such intervention, while the other showed more asymptomatic cases.

The former model did not assume the waning of vaccine-induced protection in infants after three months and showed this age shift in peak cases. It also showed that non-severe hospitalizations (not requiring admission to the intensive care unit, ICU) dropped to about a third at 3-5 months with mAb use, while those requiring primary care dropped by half.

In contrast, the other dynamic model implemented a median period of ~100 days of protection following mAbs, while herd immunity reduced transmission among mothers and children. This was also considered to avert many cases at three months to 11 months, accounting for the lower number of cases across age groups. This model, therefore, estimated higher savings for medical costs and 10-fold lower QALYs.

The differences in case estimates could be explained by the different models used, the basic assumptions about the number of non-severe cases, and the change in the efficacy of either of these interventions with case severity and over time. For instance, one model estimated a uniform ~40% efficacy against all infections, while the other applied 44% efficacy against infections leading to hospitalization. This would reflect in higher cost savings.

Overall, seasonal programs resulted in fewer hospitalizations, by 3% and 10%, with vaccination vs. mAb use, respectively, compared to year-round programs, even though the former program type involves a much smaller number of newborns. Thus, “all models estimated ICERs more in favour of seasonal programmes versus the year-round programmes from both perspectives.”

What are the implications?

When conducting a model-based RSV CEA, seasonal and catch-up programmes are best considered at the early stages of model design, especially for modelling mAb in a country with a clear seasonal RSV pattern.”

All the models except for one showed similar cost-effectiveness estimates, with either maternal vaccination or mAb administration, year-round or seasonal. The discordance with one model is mainly traceable to the different methodologies of estimating symptomatic infections that do not reach medical attention in proportion to asymptomatic infection and the failure to account for herd immunity caused by maternal vaccination.

This study is important in including static and dynamic models using waning and non-waning assumptions of efficacy. Moreover, the findings could help shape future models.

To improve the model's accuracy, future research should include age-specific incidence and prevalence rates for symptomatic and asymptomatic RSV infections among children who do not require medical care at any stage. This will require community-level studies or perhaps wastewater surveillance.

The change in QALYs due to non-severe cases exerts a major effect on the outcome of the modeling, as does the decrease in the effectiveness of either intervention over time. The impact of herd immunity in infants below six months of age is also an area to be explored. As such, these need to be carefully implemented to model such programs' cost-effectiveness.