Lancet Infectious Diseases
Mar 2018 Volume 18 Number 3 p227-356 e64-e106
Dengvaxia: age as surrogate for serostatus
Maíra Aguiar [Centre for Mathematics and Applications, Faculty of Sciences and Technology, Nova University of Lisbon, 2829-516 Caparica, Portugal email@example.com ]
Nico Stollenwerk [Centro de Matemática, Aplicações Fundamentais e Investigação Operacional, Faculdade de Ciências da Universidade de Lisboa, Lisbon, Portugal]
Published: 21 December 2017
Since April, 2016, Dengvaxia—a dengue vaccine produced by Sanofi Pasteur (Lyon, France)—has been licensed for use in 19 countries. Dengvaxia was recommended by the WHO Strategic Advisory Group of Experts (SAGE) on immunisation to be used in regions with high endemicity, as defined by a prevalence of dengue antibodies of more than 50% in the targeted age group of people aged 9–45 years.1 We have previously discussed the risks behind this vaccine recommendation,2, 3 and by analysing an age-structured model4 using the available vaccine trial data1 predicted a significant reduction in dengue virus infection-related hospital admissions if Dengvaxia is given only to individuals who are seropositive for dengue antibodies. We also predicted a significant increase in the number of dengue-related admissions, over 5 years, if the vaccine is administered without previous population screening for serostatus.4 More recently, our analysis of trial data has indicated that serostatus (and not age) is statistically discriminative with regard to the relative risk (RR) of hospital admission after vaccination, and of efficacy5—ie, the RR is similar across age classes but significantly different across serostatus groups.5
After rigorous statistical analysis of study data (taken from Martínez-Vega and colleagues’ analysis6), here we present further evidence showing that serostatus (and not age) is statistically discriminative with regard to the RR of hospital admission after Dengvaxia vaccination, and efficacy. By using the risk of dengue-related admission to hospital estimated from dengue antibody-seronegative children aged 2–8 years and infectivity measured from control groups of seropositive children younger than 9 years or seronegative individuals aged 9–16 years,6 we calculated the probability of specific numbers of individuals requiring dengue-related hospital admission in either seropositive children aged 2–8 years or seronegative children aged 9–16 years. The distribution of the number of expected admissions to hospital for younger seropositive children and older seronegative children and adolescents, and the actual number of hospital admissions for each group as reported in Martínez-Vega and colleagues’ analysis6 are shown in the appendix, showing the differences in the risk of hospital admissions between seronegative and seropositive individuals in the same age group.
The relative risk measured from seronegative individuals aged 2–8 years—the group for which Dengvaxia is not recommended—can explain and correctly estimate the observed increase of dengue-related admissions to hospital for all seronegative children enrolled during the vaccine trials, independent of age (appendix). However, it cannot explain statistically the observed number of seropositive individuals admitted to hospital of the same age (appendix).
Despite all our efforts in communicating the risks behind the Dengvaxia recommendation and its implementation without serostatus testing before vaccination,2, 3, 4, 5 an individual’s age, as a risk factor, was (until November) the only restriction for vaccine recommendation in countries with high endemicity. More than 800,000 children in the Philippines and around 300,000 individuals in Brazil were vaccinated against dengue virus in mass vaccination programmes, and many of those were possibly seronegative individuals. Communications from Sanofi in November,7 stating that vaccinated seronegative individuals are at risk of a vaccine-induced increase in the severity of a subsequent dengue virus infection, and by WHO advising that Dengvaxia be used only in people previously infected with dengue,8 are inexcusably late, since the vaccine-induced risk in seronegative individuals was already observed in the phase 3 trial data, published more than a year ago.1 Even with controversial modelling results indicating that the use of Dengvaxia in highly endemic settings could be beneficial at the population level,9 ethically no one should have been put under risk by receiving this vaccine.
We declare no competing interests.