New England Journal of Medicine
January 8, 2015 Vol. 372 No. 2
http://www.nejm.org/toc/nejm/medical-journal

Original Article
Efficacy of a Tetravalent Dengue Vaccine in Children in Latin America
Luis Villar, M.D., Gustavo Horacio Dayan, M.D., José Luis Arredondo-García, M.D., Doris Maribel Rivera, M.D., Rivaldo Cunha, M.D., Carmen Deseda, M.D., Humberto Reynales, M.D., Maria Selma Costa, M.D., Javier Osvaldo Morales-Ramírez, M.D., Gabriel Carrasquilla, M.D., Luis Carlos Rey, M.D., Reynaldo Dietze, M.D., Kleber Luz, M.D., Enrique Rivas, M.D., Maria Consuelo Miranda Montoya, M.D., Margarita Cortés Supelano, M.D., Betzana Zambrano, M.D., Edith Langevin, M.Sc., Mark Boaz, Ph.D., Nadia Tornieporth, M.D., Melanie Saville, M.B., B.S., and Fernando Noriega, M.D. for the CYD15 Study Group
N Engl J Med 2015; 372:113-123 January 8, 2015 DOI: 10.1056/NEJMoa1411037
Abstract
Background
In light of the increasing rate of dengue infections throughout the world despite vector-control measures, several dengue vaccine candidates are in development.
Methods
In a phase 3 efficacy trial of a tetravalent dengue vaccine in five Latin American countries where dengue is endemic, we randomly assigned healthy children between the ages of 9 and 16 years in a 2:1 ratio to receive three injections of recombinant, live, attenuated, tetravalent dengue vaccine (CYD-TDV) or placebo at months 0, 6, and 12 under blinded conditions. The children were then followed for 25 months. The primary outcome was vaccine efficacy against symptomatic, virologically confirmed dengue (VCD), regardless of disease severity or serotype, occurring more than 28 days after the third injection.
Results
A total of 20,869 healthy children received either vaccine or placebo. At baseline, 79.4% of an immunogenicity subgroup of 1944 children had seropositive status for one or more dengue serotypes. In the per-protocol population, there were 176 VCD cases (with 11,793 person-years at risk) in the vaccine group and 221 VCD cases (with 5809 person-years at risk) in the control group, for a vaccine efficacy of 60.8% (95% confidence interval [CI], 52.0 to 68.0). In the intention-to-treat population (those who received at least one injection), vaccine efficacy was 64.7% (95% CI, 58.7 to 69.8). Serotype-specific vaccine efficacy was 50.3% for serotype 1, 42.3% for serotype 2, 74.0% for serotype 3, and 77.7% for serotype 4. Among the severe VCD cases, 1 of 12 was in the vaccine group, for an intention-to-treat vaccine efficacy of 95.5%. Vaccine efficacy against hospitalization for dengue was 80.3%. The safety profile for the CYD-TDV vaccine was similar to that for placebo, with no marked difference in rates of adverse events.
Conclusions
The CYD-TDV dengue vaccine was efficacious against VCD and severe VCD and led to fewer hospitalizations for VCD in five Latin American countries where dengue is endemic. (Funded by Sanofi Pasteur; ClinicalTrials.gov number, NCT01374516.)

Editorial
Preventing Dengue — Is the Possibility Now a Reality?
Stephen J. Thomas, M.D.
N Engl J Med 2015; 372:172-173 January 8, 2015 DOI: 10.1056/NEJMe1413146
Dengue is a mosquito-borne flaviviral illness that is endemic in the tropics and subtropics. An estimated 390 million infections occur annually, of which 96 million have clinical manifestations.1 Although mortality is relatively lower than that for other tropical infectious diseases, the scale of human suffering and economic resources that are expended to control dengue makes it a major global public health problem.2 The factors driving transmission and infection persist without evidence of decline. For these reasons, the world needs a safe and effective dengue vaccine.

Infection with one of the four types of dengue virus (serotypes 1, 2, 3, and 4) may result in an asymptomatic infection, a mild nonspecific viral illness, classic dengue fever, or severe dengue manifested by plasma leakage, hemorrhagic tendencies, and possibly death. Patients with a second infection with a different serotype are at increased risk for severe disease. The mechanisms responsible for enhanced disease have not been completely elucidated. It is theorized the humoral and cellular convalescent immune profiles that are present after a first infection may not only fail to control a second infection with a different serotype but may also facilitate increased target-cell infection, viral replication, and generation of a so-called proinflammatory cytokine storm.3,4

The dengue-vaccine field is facing numerous challenges. First, a viable dengue vaccine must be capable of protecting against disease caused by any of the four serotypes, a process that has been burdened by the absence of a validated animal model of disease or a well-characterized human infection model. The incomplete understanding of dengue immunopathology introduces risk into clinical development programs. Finally, the reliance on neutralizing antibody assays, which are notorious for interassay variability and cross-reactivity among serotypes, to generate immunologic end-point data introduces error into data interpretation.5

After decades of attempts to develop a dengue vaccine, the results of a phase 3 efficacy trial that are now described in the Journal are a milestone. The vaccine candidate that is described by Villar et al.6 has been tested in three clinical end-point studies. In all the studies, three doses of vaccine or a control injection were administered at 0, 6, and 12 months, and all efficacy determinations were made at study month 25.

The first study was a phase 2b efficacy trial involving children between the ages of 4 and 11 years in a single center in Thailand. The trial did not meet the primary efficacy end point, with a per-protocol efficacy of 30.2%, and showed wide variation in serotype-specific efficacy: 55.6% for serotype 1, 9.2% for serotype 2, 75.3% for serotype 3, and 100% for serotype 4.7 The first phase 3 trial, which was conducted in five Asian countries and involved children between the ages of 2 and 14 years, showed a per-protocol efficacy of 56.5%, with a similar trend in serotype-specific efficacy: 50.0% for serotype 1, 35.0% for serotype 2, 78.4% for serotype 3, and 75.3% for serotype 4.8 The phase 3 trial by Villar et al., which was conducted in five Latin America countries and involved children between the ages of 9 and 16 years, had a per-protocol efficacy of 60.8%, with serotype-specific efficacies of 50.3%, 42.3%, 74.0%, and 77.7%, respectively. Additional end points included efficacy against hospitalization (80.3%) and against severe dengue (95.5%). In each of the three studies, the cohort was highly immune to at least one of the serotypes at baseline. In the phase 2b and 3 trials in Asia, average rates of seropositive status for one or more dengue serotypes were 69.5% and 67.5%, respectively; in the study by Villar et al., the average rate was 79.4%.

These studies have answered important questions with respect to the development of a dengue vaccine but have generated numerous others. Vaccine safety, immunogenicity, and efficacy were consistent across the phase 3 studies, with measures of performance similar to those in the phase 2b trial. There were no safety signals identified and no evidence of the hypothetical risk of administering a dengue vaccine to children with a mixture of seropositive and seronegative status who are living in an area in which dengue is endemic. However, it is not clear whether this favorable safety profile will be sustained through periods of waning immunity and successive dengue exposures remote from vaccination.
Vaccination of children with seropositive status produced high seroconversion rates and broad, potent neutralizing-antibody profiles. Despite such elicitation of antibody responses, why was there such disparity in efficacy across the dengue serotypes? Could too much preexisting immunity interfere with a serotype-specific vaccine response, leaving deficits in tetravalent efficacy? It is possible that the antibodies that were measured after vaccination were not all neutralizing but were a mixture of neutralizing and cross-reactive antibodies that were poorly functioning and potentially enhancing.9 If so, this could explain the discordance between the favorable serotype-specific serologic response to vaccination and the absence of corresponding serotype-specific efficacy.
Efficacy was higher in vaccine recipients with seropositive status than in those with seronegative status. Does the inferior efficacy in seronegative vaccine recipients preclude the usefulness of this vaccine for travelers or military personnel? If the vaccine is licensed and an immunization program is implemented, will this factor have an effect on its age-specific placement in the vaccination schedule?

The observed reduction in the severity of clinical disease and the prevention of hospitalization are encouraging. Although outpatient dengue has a substantial societal cost, dengue requiring hospitalization reflects morbidity.10 Is it possible that a vaccine candidate with a modest overall efficacy could be licensed and included in a national immunization program on the basis of its ability to reduce morbidity and other outcomes driving expenditures?

The efficacy trial by Villar et al. shows that we can protect populations from dengue disease and perhaps even reduce the proportion of patients with severe disease. Although the available results are not broadly generalizable across diverse populations, a foundation for additional studies has been laid. The global enrollment of more than 30,000 children in the phase 2b and 3 studies has assuaged fears focusing on the theoretical risk that dengue vaccination could predispose recipients to enhanced rates of severe disease. It remains to be seen whether licensure will be sought on the basis of these data and what effect this could have on future attempts to conduct efficacy trials with different candidate vaccines. For now, practitioners should remain optimistic that one day it will be possible to prevent dengue.