The malaria vaccine era, measured

The final verdict on the first malaria vaccine pilot is in. In Ghana, Kenya and Malawi, areas that introduced the RTS,S/AS01E vaccine into routine child immunisation saw a mortality rate ratio of 0.87 for all-cause deaths among age-eligible children, about 13% fewer deaths, compared with areas that had not yet introduced it ¹. The authors describe this as "averting about one in eight deaths" ¹. It came from a vaccine given through ordinary clinics, with incomplete coverage of the fourth dose.

Malaria vaccination has moved from efficacy trials to routine delivery, with a second vaccine, R21/Matrix-M, now made at scale. By November 2025, more than 40 million doses had been delivered to 24 African countries ⁸. This article summarises what the evidence published between September 2011 and September 2026 shows, what it does not show, what it costs, and what comes next.

The burden

Malaria remains one of the largest single causes of child death in Africa. The World Health Organization (WHO) estimates 282 million malaria cases and 610,000 malaria deaths worldwide in 2024 ⁶. The WHO African Region carried 94% of cases and 95% of deaths, and 75% of deaths in the region were in children under 5 years ⁶. WHO's fact sheet puts the African toll at 579,000 deaths ⁷. That is roughly 434,000 young children.

The deaths are concentrated. Within the WHO African Region, Nigeria accounted for 31.9% of malaria deaths, the Democratic Republic of the Congo for 11.7% and Niger for 6.1% ⁷. About two thirds of global cases and deaths occur in 11 African countries: Burkina Faso, Cameroon, the Democratic Republic of the Congo, Ghana, Mali, Mozambique, Niger, Nigeria, Sudan, Uganda and Tanzania ⁶.

The long-term trend is positive but has stalled. Since 2000, an estimated 2.3 billion malaria cases and 14 million malaria deaths have been averted ⁶. Since 2015, the global mortality rate has fallen from 14.9 to 13.8 deaths per 100,000 people at risk ⁶. But the absolute number of deaths rose from 578,000 in 2015 to 610,000 in 2024, because populations at risk grew ⁶. The 2024 mortality rate was more than three times the target set in WHO's Global Technical Strategy ⁶.

Funding is the other headwind. Total malaria funding in 2024 was an estimated US$3.9 billion, against a target of US$9.3 billion ⁶. Official development assistance fell by 21% from 2024 to 2025, largely because of reductions from the United States, and WHO reports that this disrupted surveillance and raised the risk of stock-outs and campaign delays ⁶.

This is the context for the vaccines. They arrive at a time when progress has flattened and money is tight. That makes it more important to know exactly what they deliver.

How it works

Malaria is caused by Plasmodium parasites carried by Anopheles mosquitoes. When an infected mosquito bites, it injects a form of the parasite called a sporozoite. The sporozoites travel to the liver, multiply there, and then move into the blood. Blood-stage infection causes fever, anaemia and, in severe cases, coma, breathing difficulty and death. Young children who have not yet built natural immunity are at highest risk.

Both approved vaccines, RTS,S/AS01 and R21/Matrix-M, target the sporozoite stage. They teach the immune system to make antibodies against a protein on the sporozoite surface called the circumsporozoite protein. If enough antibody is present when a child is bitten, many sporozoites are stopped before they reach or develop in the liver. In the R21 phase 3 trial, antibodies against a repeated section of this protein (the "NANP repeat") correlated with protection ³. A modelling group later found these antibody levels met the criteria for a surrogate of protection against clinical malaria ⁴.

Three practical points follow for health workers:

• Protection is partial and it wanes. The vaccines reduce the number of malaria episodes and severe cases. They do not make a child immune. Nets, prompt testing and treatment are still needed.
• The schedule matters. Protection builds over a primary series of three doses, then needs a booster. In the MVIP pilot, RTS,S was given at 6, 7, 9 and 24 months in Ghana and Kenya, and at 5, 6, 7 and 22 months in Malawi ¹. R21 was tested as three doses four weeks apart, with a booster 12 months after the third dose ³.
• Timing can be matched to the season. Where malaria comes in a short rainy-season peak, doses can be given just before the peak. This is called seasonal vaccination, and it is where both vaccines have shown their strongest results ³⁵.

WHO now recommends both vaccines for use in malaria-endemic areas ⁶.

What the evidence shows

2011–2021: protection, waning and a second vaccine

The first phase 3 RTS,S report, published in 2011, found 50.4% efficacy against clinical malaria over 14 months in the intention-to-treat analysis of the first 6,000 children enrolled at 5–17 months of age ¹⁶. The 2015 final analysis showed why follow-up matters: over a median four years, efficacy in that age group was 36.3% with a booster and 28.3% without one. Protection against severe malaria was 32.2% with a booster, but only 1.1%, with an interval spanning no effect, without it ¹⁷. Both reports came from the same trial, funded by GSK and the PATH Malaria Vaccine Initiative; they are not independent replications ¹⁶¹⁷.

A separate, smaller trial's seven-year follow-up, published in 2016, found no significant cumulative protection from three doses without a booster: efficacy was 4.4% (95% CI −17.0 to 21.9). A post-hoc analysis found rebound disease in highly exposed children ¹⁸. This does not test today's four-dose programme. It does explain why the booster, surveillance and continued use of nets and medicines matter.

R21's 2021 phase 2b trial then supplied an early signal from 450 children in Nanoro, Burkina Faso. In the higher-adjuvant group, efficacy against clinical malaria was 77% (95% CI 67–84) at six months ¹⁹. That was a single seasonal-transmission setting and a disease endpoint, not proof of a mortality reduction. The larger phase 3 and routine-delivery studies below answer different questions.

Five more recent studies extend that evidence. Two measure what happens in routine use (MVIP and EPI-MAL-003), one is the pivotal R21 efficacy trial, one is a cost-effectiveness model, and one tests the vaccine combined with seasonal malaria chemoprevention (SMC).

MVIP: fewer children died (Ghana, Kenya, Malawi)

Who and where. The Malaria Vaccine Implementation Programme (MVIP) evaluation was led by Mwapasa, Asante, Milligan, Akech, Oduro, Mathanga and colleagues, working with the ministries of health and national Expanded Programmes on Immunisation (EPI) of the three countries ¹. It was coordinated by WHO ⁶.

Design. Clusters of administrative units (districts in Ghana, sub-counties in Kenya and groups of immunisation clinics in Malawi) were randomly assigned 1:1 to introduce RTS,S in 2019 or later ¹. There were 158 clusters: 66 in Ghana, 46 in Kenya and 46 in Malawi, with 79 implementation and 79 comparison areas ¹. A network of 26,000 local reporters notified child deaths in their communities. Study staff then visited families at home to confirm details and complete a verbal autopsy ¹. The evaluation is registered as NCT03806465 ¹.

Size. By the end of 46 months, 1,289,504 children had received dose 1, 1,158,850 dose 2, 1,068,039 dose 3 and 436,527 dose 4 ¹.

Primary outcome. Mortality from any cause except injury, in children eligible for three doses. Excluding injury deaths, there were 5,576 deaths in implementation areas versus 6,152 in comparison areas among eligible children ¹. Among non-eligible children, used as a within-area reference, there were 7,534 versus 7,044 deaths ¹. The mortality rate ratio was 0.87 (95% CI 0.77–0.97; p=0.016) ¹.

Coverage. In 2022, coverage was 82.8% (95% CI 80.7–84.9) for dose 1, 71.1% (68.8–73.5) for dose 3 and 39.9% (36.9–42.9) for dose 4 ¹. The mortality benefit came despite many children missing the final dose.

Secondary outcome. WHO summarises the MVIP findings as a 13% reduction in all-cause mortality (excluding injury) and a 22% reduction in hospitalisations for severe malaria among age-eligible children ⁶. Earlier favourable evidence on safety and severe-malaria admissions from the first two years contributed to WHO's recommendation ¹.

Caveats. The paper is described as an observational evaluation of a cluster-randomised implementation, not a blinded trial ¹. The result applies to moderate coverage of three doses and low uptake of the fourth ¹.

EPI-MAL-003: severe malaria fell by more than half

Who and where. A phase 4 disease-surveillance study nested in the MVIP areas, led by Ndeketa and colleagues, at 12 sites (four per country) in Ghana, Kenya and Malawi ². It was funded by GSK ². It is registered as NCT03855995 ².

Design and size. Children younger than 18 months were enrolled in clusters with the vaccine and clusters without it, and followed by active surveillance ². The first child was enrolled on 21 March 2019. The interim analysis cut-off was 2 November 2023 ². 45,000 children were enrolled and 39,463 analysed ².

Outcomes. Compared with unvaccinated children in clusters without the vaccine, vaccinated children had country-adjusted incidence rate ratios of ²:

• 0.70 (95% CI 0.67–0.73) for any malaria
• 0.42 (0.30–0.60) for severe malaria
• 0.64 (0.56–0.72) for malaria-related hospitalisation
• 0.79 (0.74–0.84) for all-cause hospitalisation
• 0.83 (0.64–1.09; p=0.18) for all-cause mortality

In plain terms, the vaccine prevented 58% of severe malaria in routine use. Severe malaria is the main route by which malaria kills children. Among hospitalised children, the adjusted odds ratio for anaemia was 0.81 (0.73–0.90) ². A before-after comparison with children enrolled before the vaccine arrived showed similar trends ².

Caveats. This is an interim analysis over one year after the third dose ². It is observational and industry-funded. Its mortality estimate was not statistically significant (see limits below).

R21 phase 3: high efficacy, made at scale

Who and where. The phase 3 trial of R21/Matrix-M was run by the Jenner Institute at the University of Oxford, led by Datoo and colleagues, with the Serum Institute of India ³. Sites were in Burkina Faso, Mali, Kenya and Tanzania, and the author list includes West African investigators Dicko, Tinto and Ouédraogo ³. It is registered as NCT04704830 ³.

Design. A double-blind randomised trial across five sites in four countries. Children aged 5–36 months were randomised 2:1 to R21 or a control rabies vaccine ³. Half were enrolled at two sites with seasonal malaria, where doses were given before the season, and half at "standard" sites with year-round malaria, where doses followed age ³.

Size. Of 5,477 children screened, 4,878 received the first dose. 3,103 in the R21 group and 1,541 in the control group were included in the per-protocol analysis ³.

Primary outcome. Twelve-month efficacy against a first clinical malaria episode was 75% (95% CI 71–79) at seasonal sites and 68% (61–74) at standard sites ³.

Secondary outcomes. Efficacy against multiple episodes was 75% (71–78) at seasonal sites and 67% (59–73) at standard sites ³. That meant 868 (762–974) cases averted per 1,000 child-years at seasonal sites and 296 (231–362) at standard sites ³. In children aged 5–17 months, efficacy was 79% (73–84) at seasonal sites and 75% (65–83) at standard sites ³.

Safety. The most frequent adverse events were injection-site pain (18.6%) and fever (46.7%). Serious adverse events did not differ significantly between groups, and there were no treatment-related deaths ³.

Caveats. Efficacy declined modestly over the first 12 months ³. The vaccine was co-developed with its manufacturer, which also contributed funding to the linked modelling study ⁴. The trial measured cases, not deaths.

R21 modelling: what $3 a dose buys

Who. Schmit and colleagues, a modelling team working with Burkinabè trial investigators ⁴.

Design. They fitted a model linking antibody levels to protection, using three years of phase 2b data from Nanoro, Burkina Faso, then checked it against the phase 3 results ⁴. They put this inside a transmission model and projected impact over 15 years across African settings ⁴.

Results. A four-dose age-based schedule was estimated to avert 181,825 clinical cases per 100,000 fully vaccinated children in perennial settings (range 38,815–333,491) and 202,017 in seasonal settings (29,868–405,702) ⁴. At an assumed US$3 per dose, the cost was $7 per clinical case averted in perennial settings and $6 in seasonal settings ⁴. The cost per disability-adjusted life year (DALY) averted was $34 (range 29–139) in perennial and $30 (22–172) in seasonal settings ⁴. Cost-effectiveness was better where malaria was more common ⁴.

Caveats. This is a model, not a measurement. The ranges are wide.

Vaccine plus SMC: stacking the tools in the Sahel

Who and where. Dicko and colleagues, with the London School of Hygiene & Tropical Medicine and teams in Bougouni district, Mali, and Houndé district, Burkina Faso ⁵. It is registered as NCT04319380 ⁵. This paper was first published online in August 2023 and is part of the expanded evidence window.

Design. A double-blind, individually randomised trial. Children enrolled at 5–17 months were assigned to SMC plus control vaccines, RTS,S plus placebo SMC, or both, and kept receiving them until age 5 ⁵. RTS,S was given seasonally, just before the rainy season.

Size. 5,098 (94%) of the 5,433 children who completed the initial three years were re-enrolled for two more years ⁵.

Primary outcome. Over five years, clinical malaria per 1,000 person-years was 313 with SMC alone, 320 with RTS,S alone and 133 with both ⁵. The combination had a protective efficacy of 57.7% (53.3–61.7) against SMC alone and 59.0% (54.7–62.8) against RTS,S alone ⁵. RTS,S alone was non-inferior to SMC alone (hazard ratio 1.03) ⁵.

Secondary outcomes. Compared with SMC alone, the combination reduced severe-malaria admissions by 66.8% (40.3–81.5), malarial anaemia by 65.9%, blood transfusions by 68.1% and all-cause deaths by 44.5% (2.8–68.3) ⁵. No safety signals were detected ⁵.

Results at a glance

LimitsWhat didn't work (or the limits)

The effect on deaths in the case-surveillance study was not significant. In EPI-MAL-003, the all-cause mortality rate ratio was 0.83, with a confidence interval of 0.64–1.09 that includes no effect ². The larger, cluster-randomised MVIP evaluation is the stronger result on deaths ¹, but readers should know the smaller study did not confirm it on its own.

The fourth dose is often missed. Only 39.9% of children got dose 4 in 2022 ¹. Across the whole programme, 436,527 children got dose 4 against 1,289,504 who got dose 1 ¹. The mortality benefit was measured at that level of coverage. The evidence does not tell us how much more would be gained at full coverage, but it shows that schedule completion is a weak point.

Efficacy is lower where malaria is year-round. R21's efficacy was 68% at standard sites compared with 75% at seasonal sites, and cases averted per 1,000 child-years were about a third as high ³. Seasonal delivery fits some settings better than others.

Protection wanes. The R21 trial saw a modest decline in efficacy over the first 12 months ³. A 2026 modelling study by McCain and colleagues notes that waning efficacy and the small vaccine-eligible age group mean routine vaccination is unlikely to prevent severe cases in older children, who still carry significant burden ⁹.

The mortality estimate for vaccine plus SMC is imprecise. The 44.5% reduction in deaths has a confidence interval from 2.8% to 68.3% ⁵. Deaths excluding external causes or surgery fell by 41.1%, with a confidence interval from −9.2% to 68.3% that includes no effect ⁵. Deaths from malaria fell by 66.8% (−2.7% to 89.3%), also not significant ⁵. It is a real signal, but the true size could be small or large.

Industry involvement. EPI-MAL-003 was funded by GSK ². The R21 vaccine was co-developed with the Serum Institute of India, which also helped fund the modelling study ⁴⁸. The MVIP result is the most independent of the five.

No head-to-head comparison. None of the studies here compared RTS,S and R21 directly. Differences in their headline numbers reflect different designs, settings and outcomes.

Cost and scale

Price per dose. Under a deal announced on 24 November 2025 by Gavi, the Vaccine Alliance, and UNICEF with the Serum Institute of India, the price of R21 is expected to fall to US$2.99 per dose, taking effect about one year after the announcement ⁸. The deal is financed by Gavi through an advance payment enabled by the International Finance Facility for Immunisation (IFFIm) ⁸. UNICEF expects savings of up to US$90 million, equal to more than 30 million additional doses and nearly 7 million more children vaccinated over five years ⁸.

Cost per health gain. At US$3 per dose, modelling puts R21 at about $6–7 per clinical case averted and $30–34 per DALY averted ⁴. That places the vaccine among the best-value tools in child health. The R21 modelling results quoted here are per case and per DALY averted; we do not convert them into a cost per death averted.

Coverage and rollout. By the end of 2024, malaria vaccines had been introduced in 17 countries, with UNICEF delivering over 10.5 million doses and enabling vaccination of at least 2.1 million children during the year ⁶. By October 2025, seven more countries had introduced the vaccine ⁶. UNICEF reports that as of November 2025 more than 40 million doses had been delivered to 24 African countries, with Burundi, Uganda, Mali, Guinea, Togo, Ethiopia and Zambia introducing in 2025 ⁸. The MVIP paper notes that malaria vaccines have been added to immunisation schedules in 25 sub-Saharan African countries ¹.

Who funds and delivers. Gavi co-finances vaccine purchase and supports country introductions, UNICEF procures and ships the doses, and national EPIs within ministries of health deliver them through routine clinics ⁶⁸. In the MVIP pilot, delivery ran through the national immunisation programmes of Ghana, Kenya and Malawi ¹.

The funding squeeze. Vaccines sit inside a malaria budget that reached only 42% of the required amount in 2024 ⁶. In 2024, about 44% of malaria funding came from endemic countries themselves, up from a 2010–2024 average of 33% ⁶.

What to watch

Vaccine synchronised with SMC, at scale. Epicentre is running two phase 4 cluster-randomised trials. One, in Burkina Faso and Mali (NCT06860178, 40,000 planned), compares routine R21 vaccination with an annual three-dose campaign aligned with SMC distribution for children aged 3–59 months; the primary outcome is malaria incidence at 12 months ¹⁰. The other, in Moïssala and Dembo districts of Chad (NCT07038837, 70,000 planned), randomises 26 health-centre catchment areas to routine R21 or R21 synchronised with the SMC campaign ¹¹. Both are recruiting, with primary completion listed for 2026 ¹⁰¹¹. They will test whether the Dicko trial's result ⁵ holds in routine programmes.

R21 in routine use. The AVERT study, led by the Clinton Health Access Initiative (NCT07749911), is a test-negative case-control study of R21 effectiveness against clinical malaria in Burkina Faso and Uganda, recruiting with 20,000 planned ¹². The Serum Institute of India is running an observational effectiveness and safety study in Burkina Faso, Côte d'Ivoire, Nigeria and Uganda (NCT07009847) ¹³. These are the R21 equivalents of EPI-MAL-003.

Vaccine plus perennial chemoprevention. Stanford University plans a trial at Masafu General Hospital in Busia District, Uganda (NCT07322068, not yet recruiting), testing whether adding perennial malaria chemoprevention with SPAQ or dihydroartemisinin-piperaquine to R21 improves protection over R21 alone ¹⁴. The registry entry notes that R21 does not protect until the primary series is complete, at about 9 months of age ¹⁴.

Next-generation vaccines. The University of Oxford is testing multi-stage combinations of R21 with the blood-stage candidates RH5.1 and R78C in Burkina Faso (NCT07183371) ¹⁵.

Older children and boosters. McCain and colleagues estimate that, in moderate-to-high transmission, catch-up campaigns for children under 5 or extra boosters would prevent more cases per dose than routine vaccination in low-transmission areas ⁹. For example, an extra booster 5 years after the primary series averted 835 clinical cases per 1,000 additional doses at 45% parasite prevalence, against 247 per 1,000 doses for routine vaccination at 5% prevalence ⁹. The authors assume older children respond as well as infants, which trials have not confirmed ⁹. This is an open question for guidance.

What it means

For ministries of health. A malaria vaccine delivered through routine immunisation reduces child deaths, not only malaria cases ¹. It halves severe malaria in ordinary clinics ². The biggest gain within reach is the fourth dose: fewer than four in ten children received it in the pilot ¹. Linking the booster to other contacts at 18–24 months, tracking defaulters, and counting dose 4 in EPI performance reviews are practical steps.

For programme planners in the Sahel and other seasonal settings. The best documented approach is to combine tools: seasonal vaccination plus SMC gave fewer cases, fewer severe admissions and fewer deaths than either tool alone ⁵. The Epicentre trials in Burkina Faso, Mali and Chad will show whether this works through routine campaigns ¹⁰¹¹.

For funders. The cost is low: about $30–34 per DALY averted at $3 a dose ⁴, and the R21 price is set to fall to $2.99 ⁸. R21's efficacy and scale of manufacture explain why rollout can expand beyond the pilot countries ³. With total malaria funding at US$3.9 billion against a US$9.3 billion need ⁶, vaccines should be protected alongside nets and chemoprevention, not traded against them.

For clinicians and health workers. A vaccinated child can still get malaria. Test every fever and treat promptly. Keep children under nets. Remind caregivers that the booster dose counts: every child who completes the schedule adds to the protection that has already been shown to save lives.