In early 2025, Sri Lanka's health ministry reported a troubling trend: despite a national vaccination campaign that had reached roughly 70% of children in high-risk districts, hospitalised dengue cases rose by nearly a third compared to the same period the previous year. The vaccine in question, Dengvaxia, had been rolled out with cautious optimism since 2017, following a phase 3 trial that showed about 60% efficacy against symptomatic dengue. But the virus, it seemed, was adapting faster than the immunity the vaccine could provide.

The discord among experts is not merely academic. It has real consequences: a 15% drop in vaccination uptake in Colombo's suburban clinics was reported in early 2026, as parents began questioning the vaccine's value. At the centre of the debate is a simple but vexing question: does Dengvaxia still protect enough to justify its use, given the shifting serotype landscape?

A vaccine that worked in trials, then stumbled in the field

Dengvaxia, developed by Sanofi Pasteur, was first licensed in 2015 after a large phase 3 trial across Asia and Latin America demonstrated an overall efficacy of roughly 60% against symptomatic dengue, with higher protection against severe disease. The vaccine was designed to protect against all four dengue serotypes, but its performance varied. In seropositive individuals, those previously infected with dengue, efficacy was higher; in seronegative individuals, the vaccine paradoxically increased the risk of severe dengue upon subsequent infection, a phenomenon known as antibody-dependent enhancement (ADE).

Despite these caveats, Sri Lanka's National Immunization Technical Advisory Group (NITAG) recommended a targeted vaccination program for children aged 9 to 16 in high-incidence districts, after a pre-vaccination screening test to ensure only seropositive children received the vaccine. The program began in 2018, and by 2023, coverage in the target population was estimated at roughly 65–75%, according to ministry data.

But by 2024, hospital surveillance data began to show a worrying pattern. The proportion of hospitalised dengue cases that were fully vaccinated rose from about 15% in 2020 to nearly 30% in 2024. While this could partly reflect high coverage, the absolute number of breakthrough infections also increased. A post-licensure observational study, published in early 2025, found that vaccine efficacy against serotype 2 had dropped to roughly 35% in the 2024 season, down from an estimated 50% in earlier years. The WHO prequalification status of Dengvaxia remains unchanged, and global guidelines still recommend its use in seropositive populations. However, the Sri Lankan NITAG is now split. Some members argue that the vaccine still reduces severe disease by about 50% overall, and that stopping the program would leave children vulnerable. Others counter that the number needed to vaccinate to prevent one hospitalisation has climbed above an acceptable threshold, and that the program should be paused pending a review of new candidate vaccines.

Why serotype mismatch matters more than aggregate efficacy

Dengue is caused by four distinct serotypes of the virus. Infection with one serotype confers lifelong immunity to that serotype but only partial and transient protection against the others. Subsequent infection with a different serotype increases the risk of severe dengue, due to ADE. An effective vaccine must therefore induce balanced, durable immunity against all four serotypes.

Dengvaxia is a chimeric vaccine based on a yellow fever backbone, expressing the prM and E proteins of each dengue serotype. Early trials showed that antibody responses were strongest against serotypes 3 and 4, but weaker against serotypes 1 and 2. Over time, antibody titres against serotype 1 waned fastest, according to data from the phase 3 trial extension. In Sri Lanka, the serotype distribution has shifted markedly in recent years. During the 2017 outbreak, serotype 2 accounted for roughly 60% of cases. By 2024–2025, serotypes 2 and 3 together made up about 70% of sequenced isolates, with serotype 1 declining to less than 15%. This mismatch—the vaccine's weaker coverage against the dominant serotypes—helps explain the declining efficacy. The risk of ADE in vaccinated individuals who later become infected remains a theoretical concern, but it has been difficult to measure directly. In the Philippines, a 2017 suspension of Dengvaxia after reports of increased severe disease in seronegative recipients led to a crisis of confidence. Sri Lanka's pre-vaccination screening was designed to avoid that scenario, but some experts worry that the screening test's sensitivity is imperfect, and that some seronegative children may have been vaccinated inadvertently.

Two camps of epidemiologists, one set of national data

The debate within Sri Lanka's epidemiological community has crystallised into two camps. The first camp, which includes several members of the NITAG, argues that the vaccine still provides meaningful protection against severe disease, hospitalisation, and death. They point to hospital surveillance data showing that vaccinated children who do get dengue are less likely to develop dengue haemorrhagic fever or require intensive care. A 2025 analysis by the Ministry of Health estimated that the vaccine prevented roughly 2,000 hospitalisations and 50 deaths in 2024 alone, based on a comparison of observed cases in vaccinated versus unvaccinated cohorts.

The second camp counters that these estimates are inflated by confounding. They note that vaccinated children are more likely to be from urban areas with better access to healthcare, and that the unvaccinated comparison group includes many children who were screened seronegative and thus at lower baseline risk of severe disease. When adjusted for socioeconomic status and prior infection, the vaccine's effectiveness against hospitalisation drops to roughly 40%, they argue. The number needed to vaccinate to prevent one hospitalisation has risen from about 50 in 2019 to over 200 in 2024, according to their calculations.

Both camps cite the same national hospital surveillance system, but they emphasise different endpoints. The first camp focuses on severe dengue and death, where the vaccine appears more effective. The second camp focuses on all hospitalised dengue cases, where the effect is weaker. The Ministry of Health convened a technical working group in March 2026 to try to reconcile these analyses, but no consensus has emerged.

The working group's terms of reference include evaluating the potential role of TAK-003, a newer dengue vaccine developed by Takeda that has shown more balanced efficacy across serotypes in phase 3 trials. However, TAK-003 has not yet been licensed in Sri Lanka, and the cost of switching would be substantial. Some experts argue that the country should wait for more data, while others see the delay as a missed opportunity. TAK-003, also known as Qdenga, demonstrated overall efficacy of about 80% against hospitalised dengue in a phase 3 trial across Latin America and Asia, with similar protection across serotypes 1, 2, and 3. However, efficacy against serotype 4 was lower, and the vaccine is not recommended for seronegative individuals in some settings. A modelling study by the University of Colombo estimated that replacing Dengvaxia with TAK-003 could prevent an additional 1,500 hospitalisations per year, but at an incremental cost of roughly US$ 3–5 million annually, a sum that would require reallocation from other health programs.

The role of vector control in muddying the evidence

Adding to the confusion is the fact that vector control efforts have been uneven across Sri Lanka. In Colombo, a large-scale fogging campaign in 2023 reduced the Aedes aegypti mosquito population by roughly 40%, according to entomological surveys. This reduction likely lowered dengue transmission in the capital, making the vaccine appear more effective there than in districts where vector control was less successful.

In contrast, outer districts such as Jaffna and Batticaloa saw little improvement in vector indices. In these areas, the incidence of dengue remained high, and vaccine effectiveness estimates were lower. Observational studies that attempt to adjust for vector density are limited by the quality of entomological data, which is collected sporadically and not standardised across districts.

The problem is that vaccine effectiveness is typically measured by comparing disease rates in vaccinated and unvaccinated individuals. If the risk of exposure differs between groups—because of differences in mosquito density, for example—then the comparison is biased. Cluster-randomised trials, which randomise entire communities to vaccine or control, can avoid this bias, but they are expensive and logistically complex. A proposed cluster-randomised trial in Sri Lanka, designed to estimate vaccine effectiveness under real-world conditions, has been stalled due to lack of funding.

The debate over how to interpret the Sri Lanka data echoes earlier controversies over rotavirus vaccine in India, where post-licensure effectiveness was lower than expected due to differences in circulating strains and underlying health status. In both cases, the tension between global recommendations and local evidence has been difficult to resolve.

What clinicians in Colombo and Jaffna are seeing on the ward

At Lady Ridgeway Hospital in Colombo, the paediatric dengue ward has been near capacity for most of the 2025–2026 season. Dr. Anusha Perera, a senior paediatrician who has worked in the ward for over a decade, says she has noticed a change in the clinical presentation of dengue in vaccinated children. “They often present with atypical symptoms—less fever, more abdominal pain and vomiting,” she explains. “The classic warning signs of dengue, like mucosal bleeding, are sometimes absent. It makes diagnosis more challenging.”

Physicians report an increase in thrombocytopenia, or low platelet count, among vaccinated children compared to pre-vaccine years. While severe bleeding remains rare, the trend has prompted many clinicians to order serotype-specific PCR tests for all suspected dengue cases, a practice that was previously reserved for research settings. The tests are not covered by the public health system, so families must pay out of pocket—roughly 5,000 Sri Lankan rupees, a significant sum for many.

In Jaffna, the situation is different. The Teaching Hospital in Jaffna has seen a surge in dengue cases this season, but the proportion of vaccinated children among them is lower than in Colombo, partly because vaccination coverage in the Northern Province is only about 50%. Dr. Sivapalan Sivakumar, a consultant physician, says his team has not observed a clear difference in clinical presentation between vaccinated and unvaccinated patients. “We are still seeing typical dengue fever with rash and myalgia,” he notes. “But we don't have routine serotyping here, so we can't say for sure.”

Local treatment protocols have been adapted informally. In some hospitals, clinicians have begun using a lower threshold for platelet transfusion in vaccinated children, despite a lack of national guidance. The Ministry of Health has not issued updated clinical guidelines since 2022, leaving individual practitioners to make decisions based on their own experience and interpretation of the evidence.

Pressure from WHO and global funders to stay the course

The World Health Organization's Global Advisory Committee on Vaccine Safety reviewed the Sri Lanka data in April 2026 and issued a statement that found “no evidence of harm” from the vaccine, but called for “continued monitoring of efficacy and serotype distribution.” The statement was careful not to recommend changes to the vaccination program, but it also did not endorse the current strategy unequivocally.

Gavi, the Vaccine Alliance, which has provided financial support for Sri Lanka's dengue vaccination program, committed to its 2027 procurement targets for Dengvaxia in a board meeting in May 2026. A Gavi spokesperson said the organisation is “closely following the situation” but deferred to national authorities on policy decisions. Critics argue that global bodies are slow to update serotype-specific recommendations, and that the incentives of large organisations favour continuity over change.

The parallels to the 2017 Philippines suspension of Dengvaxia are often cited by both camps. In the Philippines, the vaccine was withdrawn after reports of increased severe disease in seronegative children, leading to a national scandal and a sharp decline in vaccine confidence. Some experts in Sri Lanka worry that a similar loss of trust could occur if the program is perceived as failing, even if the risk profile is different.

Dr. Nimali Abeysekera, an epidemiologist at the University of Colombo who sits on the NITAG, argues that the global pressure to maintain the program is driven by a desire to protect the vaccine's reputation. “We have invested so much in Dengvaxia, both financially and in terms of public trust,” she says. “Admitting that it is not working as well as hoped is difficult for everyone involved.”

What a credible serotype-surveillance system would require

One thing both camps agree on is that the current surveillance system is inadequate for resolving the debate. Sri Lanka's national dengue surveillance is based on a network of sentinel hospitals that report cases weekly, but only a fraction of samples are serotyped. In 2025, only about 5% of hospitalised cases had a serotype result recorded, and those were concentrated in Colombo and Kandy.

A credible serotype-surveillance system would require sentinel sites in each province with PCR capacity year-round, not just during outbreaks. Samples would need to be collected systematically from a representative subset of cases, and results reported in real time to a central dashboard that could be accessed by researchers and policymakers. The cost of such a system is estimated at roughly US$ 2–4 million annually, a significant but not prohibitive sum for a country with a health budget of about US$ 2 billion.

Linking the vaccine registry to hospital discharge data would allow for more accurate estimates of vaccine effectiveness by serotype. Currently, the two databases are not interoperable, so linking requires manual matching of names and dates of birth—a time-consuming process that is prone to error. A unique health identifier, which has been discussed for years, has not yet been implemented.

Without such a system, experts say the debate will remain unresolved. “We are essentially arguing over the same incomplete data,” says Dr. Abeysekera. “We need better data, not more analyses of the same data. But better data costs money, and that money is not available.” The question now is whether Sri Lanka can afford to wait for perfect data before deciding the future of its dengue vaccination program, or whether the rising hospitalisations will force a decision sooner.