Dengue vaccine safety update

The WHO Global Advisory Committee on Vaccine Safety has been following the development of a tetravalent recombinant live dengue virus vaccine for the past 5 years.^{9, 10, 11, 12} The most advanced product, CYD-TDV by Sanofi-Pasteur (Dengvaxia®) is a yellow fever vaccine virus backbone vector that expresses envelope proteins of dengue viruses type 1 to 4. During early clinical trials, no serious vaccine-related events had been documented among the recipients, and no excess cases of dengue fever or severe dengue attributable to the vaccine had been observed. Subsequent large-scale phase 3 trials, CYD14 in Asia (among subjects aged 2–14 years) and CYD15 in Latin America (among subjects aged 9–16 years) were conducted in over 20 000 vaccine recipients and 10 000 control subjects and demonstrated partial efficacy of the vaccine.

Dengvaxia® received its first marketing authorizations in late 2015 and is currently available in several Asian and Latin American countries. This report briefly reviews the experience presented to GACVS during clinical trial development (now with >5 years of follow up), and discusses new evidence presented to WHO during the GACVS meeting in December 2017. These new data are based on the reanalysis of clinical trial data using a new test that retrospectively distinguished subjects with and without prior exposure to wild dengue virus.

Background

Dengue is an increasingly important disease worldwide. As outlined in the WHO position paper on dengue vaccines published in July 2016,¹³ the number of cases reported annually to WHO increased from 0.4 to 1.3 million during the decade 1996–2005, reaching 2.2 million in 2010 and 3.2 million in 2015. Based on mathematical modelling, the global annual incidence has been estimated at approximately 50–100 million symptomatic cases, predominantly in Asia, followed by Latin America and Africa. However clinical cases are likely to represent only about 25% of all dengue virus infections. In 2013 dengue was estimated to be responsible for approximately 3.2 million severe cases and 9000 deaths, the majority occurring in lower middle-income countries. GACVS recognized several challenges for the evaluation of the safety of dengue vaccines, particularly the follow-up time needed to monitor the theoretical risk of increased severe dengue following vaccination. Severe dengue cases represent a small percentage of all dengue infections and are more common on second exposure to wild dengue viruses.

In 2015, the Committee was presented with results from the third year of follow-up in the CYD14 trial conducted in Asia. The trial found that the risk of hospitalized dengue was significantly higher in the vaccinated group compared with the control group of age 2–5-years (relative risk (RR) = 7.45, 95% confidence interval (CI): 1.15, 313.80). This risk was not found to be elevated in older age groups. At the time, GACVS highlighted the importance of understanding potential factors, other than age, that may be associated with this increased relative risk of hospitalization and of severe dengue. Among them, understanding if a subject had been exposed to wild dengue virus prior to vaccination was deemed critical given the lower vaccine efficacy in participants who were serologically naïve, and the potential risk of immune enhancement among previously infected subjects.

Based on these results, GACVS noted that the excess cases of hospitalized dengue (in the age group 2–5 years) could be related to age, serostatus, or both. The plausible hypothesis proposed was that vaccination primes the immune system similarly to natural infection, and that after a period of protection following vaccination, immunity wanes. According to this hypothesis, among seronegative individuals, the response to the first natural infection following vaccination (and waning immunity) may act as a second infection, which has typically been associated with a higher risk of serious disease. In seropositive individuals, the response to the first natural infection following vaccination is as if it was a third or later infection and not associated with a higher risk of serious disease. As a result of available evidence, licensure was sought for children and adults aged ≥9 years. The Strategic Advisory Group of Experts (SAGE) on Immunization issued recommendations in April 2016 to introduce Dengvaxia® in geographical settings (national or subnational) with high endemicity only, as indicated by seroprevalence of >70% in the age group targeted for vaccination.

In June 2016,⁴ GACVS was presented with the longer-term 4-year follow-up of hospitalized dengue among CYD14 and CYD15 clinical trial participants. While no consistent increase was observed in the risk of hospitalization or severe dengue in vaccinated individuals aged 9–16 years, in the younger age group of 2–8 years, an increased relative risk (not reaching significance) was observed in year 3 of follow-up that persisted during year 4 but was declining. GACVS recommended that existing and planned clinical efficacy trials should be evaluated in depth and include careful assessment of pre-immunization seropositivity in selected cohorts. These data would contribute to a greater understanding of the potential risk factors and underlying immunology of dengue infection and severe dengue post-vaccination.

Current status and new data

To date, the vaccine has been licensed in 19 countries and introduced in public immunization programmes in the Philippines and Brazil. Immunization began in the Philippines in April 2016 and GACVS was presented with the programme’s early post-market surveillance experience.⁴ The country had seen dramatic increases in cases since 2010 with >150 000 dengue episodes and approximately 1000 deaths annually. By the time of the meeting in June 2016,⁴ almost 250 000 children aged ≥9 years had been vaccinated.

As SAGE identified vaccine safety in the seronegative population as a research priority,⁶ Sanofi Pasteur has undertaken a case-cohort study using a dengue anti-NS1 IgG ELISA assay (NS1) on blood samples available from clinical trial participants at 13 months after the first dose (1 month following the third and last dose of vaccine administered during the clinical trials). The research assay is designed to differentiate between prior natural infection and vaccination. Based on these results, the company reanalysed the safety and efficacy according to this surrogate of serostatus as well as age at the time of vaccination.

Overall, vaccinated trial participants had a reduced risk of virologically-confirmed severe dengue and hospitalizations. The subset of trial participants who had not been exposed to dengue virus infection prior to vaccination (i.e. dengue-naïve, seronegative according to the NS1 assay) had a twice higher risk of more severe dengue and hospitalizations compared with unvaccinated participants, regardless of age. In contrast, those trial subjects, at any age, with evidence of a previous dengue infection (as determined by NS1 assay) experienced a reduced risk of severe dengue for the duration of the observation period.

Based on this new analysis, Sanofi Pasteur estimated the actual risks in the study population. In study subjects aged 2–16 years without prior dengue infection, data suggest modest efficacy (15–32%) of vaccine against symptomatic dengue until the second year of follow-up. Subsequently, starting during year 3, the risk of hospitalized and severe illness becomes higher than among controls. In practical terms, and within the population studied, these data suggest that during a 5-year follow-up, approximately 5 additional hospitalized dengue cases, or 2 additional severe dengue cases, per 1000 vaccinees with no previous dengue infection (i.e. dengue naïve subjects) could occur following vaccination, compared with unvaccinated seronegative children. Importantly, in the clinical trial population, all cases recovered and no deaths were observed. On the other hand, among children who had a prior dengue infection (i.e. seropositive) there was a reduction of 15 cases of hospitalized dengue and 4 cases of severe dengue per 1000 who were vaccinated for the same duration of follow-up.

Implications and assessment by GACVS

GACVS considered these new results as well as the clinical trial data and early post-market data submitted. Key issues included the validity of these findings, the subgroups to which they mostly apply, the magnitude of the risk, and implications, both for those subjects already vaccinated, and those not yet vaccinated. GACVS acknowledged that the vaccine is safe and efficacious in individuals who have had a primary infection with wild dengue preceding immunization, thus preventing a “second” and therefore more severe episode of dengue. GACVS noted that the increased risk of severe dengue among vaccinated individuals who are seronegative to dengue at the time of vaccination became apparent during the third year after receipt of the first vaccine dose irrespective of age. Thus, as post-licensure use started in the second quarter of 2016, an increase in the number of severe dengue cases among seronegative subjects would not occur before 2018 in Brazil and the Philippines – the 2 countries with early introduction of the vaccine.

The new data indicate that the increased risk of hospitalization (and severe disease) from dengue affects vaccinated subjects who are naïve to wild dengue infection prior to vaccination. This provides strong indication that previously identified excess risk among younger vaccine recipients in the Asian study reflected a confounding association between age and exposure to wild dengue virus. Thus it appears that history of exposure to wild dengue, rather than age, predicts the risk of severe disease among vaccine recipients. This also corroborates prior hypotheses suggesting that immune priming from natural or other stimulation such as immunization with the dengue vaccine can lead to a higher risk of severe dengue disease on secondary exposure to wild dengue viruses.

GACVS recognizes that the vaccine has, to date, been administered to a large majority of subjects among populations where exposure to dengue virus is high and therefore most vaccine recipients are seropositive to wild dengue. Notable is that the clinical data presented by Sanofi Pasteur also showed that, even among seronegative population, the number that would experience untoward severe dengue is likely to be <1%, and that with proper clinical care, more serious consequences can be prevented in most instances.

As a result, GACVS recommends that Dengvaxia® should not be administered to individuals who have not been previously infected with wild dengue virus. Data are not currently available to allow an analysis of the risk according to the number of vaccine doses received by subjects seronegative at baseline. It is therefore not possible to determine if incomplete vaccination would lead seronegative subjects to a higher or lower risk of severe dengue as compared to seronegative subjects who have received the full 3-dose course.

In order to minimize untoward consequences for dengue-naïve vaccinated subjects, GACVS recommends ensuring the enhancement of measures that reduce exposure to dengue infection among populations where the vaccine has already been administered. For vaccine recipients who present with clinical symptoms compatible with dengue virus infection, access to medical care should be expedited to allow for proper evaluation, identification, and management of severe forms of the disease.

⁴ See No. 28, 2017, pp. 393–396.

⁶ See No. 7, 2014, pp. 57–58.

⁹ See No. 06, 2013, pp. 68–69.

¹⁰ See No. 04, 2015, pp. 17–18.

¹¹ See No. 34, 2015, pp. 421–423.

¹² See No. 28/29, 2016, pp. 346–347.

¹³ See No. 30, 2016, pp. 349–364.

Full report of GACVS meeting of 6-7 December 2017, published in the WHO Weekly Epidemiological Report on 19 January 2018

GACVS last reviewed dengue vaccines in June 2015⁵ and considered the Phase III clinical trial and long-term safety data of the tetravalent dengue vaccine CYD-TDV (Dengvaxia, manufactured by Sanofi Pasteur). Short-term safety surveillance of common adverse events demonstrated that the vaccine is well-tolerated. However, GACVS noted particular safety concerns related to the observed risk of hospitalized and severe dengue among children aged 2–5 years during the third year following vaccination. GACVS also recommended monitoring the risk of severe dengue among individuals of all ages who are seronegative prior to immunization, as well as among immunocompromised and older individuals (>45 years of age).

In April 2016, SAGE published recommendations indicating who would benefit most from CYD-TDV vaccination and issued guidelines for post-licensure surveillance.⁶ In particular, SAGE recommended that countries considered introducing CYD-TDV only in geographical settings (national or subnational) with high endemicity, as indicated by seroprevalence of approximately >70% in the age group targeted for vaccination or other suitable epidemiologic markers. Pregnancy remains a contraindication.

To date, although licenced in several countries, CYD-TDV has been introduced to the public vaccination programmes of one country alone – the Philippines. Here, CYD-TDV vaccination is administered as part of a school-based programme targeting 4th grade children (aged 9–10 years) in 3 highly endemic regions. The first child in the Philippines was vaccinated in April 2016. Currently, approximately 247 820 children are immunized, with a planned total cohort of about 750 000. In this cohort AEFI are monitored through enhanced passive surveillance. All serious AEFI are investigated promptly and reviewed by an independent expert committee. Thus far, 518 AEFIs have been reported including 21 serious AEFIs with 2 deaths. Those included 2 anxiety reactions and 4 cases classified as consistent with a causal association to the vaccine that recovered.

GACVS noted the commitment of the Government of the Philippines, the community and the vaccine and health-care providers in implementing the dengue vaccination programme. The Philippines, as the first country to introduce CYD-TDV vaccination, will thus provide critical post-licensure AEFI surveillance data that will benefit and inform the global community and other jurisdictions.

Representatives of Sanofi Pasteur, the manufacturer of the CYD-TDV vaccine Dengvaxia, gave a presentation to GACVS in which they addressed the longer-term follow-up of hospitalized dengue among Phase III clinical trial participants. With follow-up occurring after 4 years since the first dose of vaccine, no consistent increase was observed in the relative risk of hospitalization or severe dengue in vaccinated individuals aged 9–16 years. However in the younger age group of 2–8 years, an increased relative risk (RR>1, not reaching significance) was observed that declined after 3 years since the first dose. Dengvaxia is not licensed for children aged <9 years.

Following the introduction of dengue vaccination programmes, GACVS recommends robust, ongoing surveillance with particular emphasis on establishing disease and vaccination history. This requires allocating resources specifically to vaccination registries and ensuring that cases of hospitalized dengue are confirmed in accordance with established case definition. This may be feasible at sentinel sites only. Existing and planned clinical efficacy trials should be evaluated in depth and include careful assessment of pre-immunization seropositivity in selected cohorts. Data from these trials will contribute to a greater understanding of the potential risk factors and underlying immunology of dengue infection and severe dengue post-vaccination.

⁵ See No. 34, 2015, pp. 421–423.

⁶ See No. 21, 2016, pp. 282–284.

Full report of GACVS meeting of 15-16 June 2016, published in the WHO Weekly Epidemiological Record of 15 July 2016