Bivalent RSV prefusion vaccine effectiveness against hospitalisation in older adults: meta-analysis of case–control studies in England, Wales, Scotland, and Northern Ireland

Summary

Background

Respiratory syncytial virus (RSV) causes substantial morbidity and mortality in older adults. In 2023, the RSVpreF vaccine was licenced and recommended for adults through the UK's national immunisation programmes. Real-world evidence on vaccine effectiveness (VE) across different populations and healthcare settings is limited.

Methods

We conducted a retrospective, multi-nation, test-negative design analysis to evaluate RSVpreF VE against hospitalisation in adults aged 74–79 years across England, Wales, Scotland, and Northern Ireland during the 2024–25 RSV season. Laboratory testing, admission, and vaccination data were linked at the individual patient level. Eligible cases were hospitalised individuals with laboratory-confirmed RSV, and controls were RSV-negative hospitalised patients; SARS-CoV-2- and influenza-positive controls were excluded in the primary analysis. Nation-specific VE estimates were derived using multivariable logistic regression, adjusted for epidemiological week of specimen collection, and combined using fixed-effects meta-analysis with inverse-variance weighting. Sensitivity analyses included redefinition of controls and leave-one-nation-out analyses.

Findings

A total of 11,117 adults were included (6528 in England, 942 in NI, 1462 in Scotland, 2185 in Wales). Overall, 3896 (35.0%) had received RSVpreF. Across nations, 826 RSV-positive cases were identified, of whom 726 (87.9%) were unvaccinated. Adjusted VE against RSV-related hospitalisation was 74% (95% confidence interval [CI] 68–80) in England, 77% (30–95) in NI, 81% (60–91) in Scotland, and 82% (43–94) in Wales. Pooled fixed-effects VE was 75% (69–80; I² = 0%). VE estimates from sensitivity analyses were robust to the inclusion of vaccine-preventable respiratory infection controls and adjustment for vaccination status. Leave-one-nation-out sensitivity analysis showed that excluding England increased pooled VE to 81% (66–89), whereas exclusion of NI, Scotland, or Wales had minimal impact (75%, 68–80 to 69–80).

Interpretation

RSVpreF vaccination provides high real-world protection against RSV-related hospitalisation in adults across the UK.

Funding

This study received no specific funding.

Research in context.

Evidence before this study

In 2023, a large multinational phase 3 clinical trial demonstrated that the bivalent prefusion F (RSVpreF) vaccine (Abrysvo®, Pfizer) was safe and effective in preventing RSV-associated lower respiratory tract infection in older adults. National immunisation recommendations for older adults have since been introduced in a number of countries, including the United States, United Kingdom (UK), Canada, Germany, Denmark, Austria, Switzerland, France, Italy, Belgium, and Sweden. In the UK, the vaccine was introduced in August–September 2024, and offered to individuals aged 75 years, with a catch-up campaign targeting those aged 76–79 years. Eligibility criteria were broadly consistent across the four nations, although minor variations in implementation exist, including differences in programme start dates and local delivery arrangements.

To assess the existing evidence on real-world VE of RSVpreF, we searched PubMed on October 1, 2025, using the terms (“RSV” OR “respiratory syncytial virus”) AND “vaccine effectiveness” AND (“Abrysvo” OR “RSVpreF” OR “Pfizer”), with no date or language restrictions. The search returned 53 articles. Since existing post-licensure studies are almost entirely US-based, we manually excluded US studies to focus on evidence from other settings. Only one relevant study was identified: a pragmatic, open-label trial from Denmark that enrolled adults aged ≥60, which reported a vaccine effectiveness of 83.3% (95% CI: 42.9–96.9) against hospitalisation for RSV-related respiratory disease.

Added value of this study

This study provides the real-world evaluation of RSVpreF (Abrysvo®, Pfizer) vaccine effectiveness against RSV-related hospitalisation in older adults during the first RSV season following vaccine implementation in the UK. Using meta-analysis of national test-negative case–control studies across England, Northern Ireland, Scotland, and Wales, including 11,117 adults aged 75–79 years, we observed high and consistent protection, with a pooled adjusted vaccine effectiveness of 75% (69–80). Sensitivity analyses confirmed the robustness of these findings across nations and analytical approaches.

Implications of all the available evidence

This UK analysis demonstrates that RSVpreF (Abrysvo®, Pfizer) vaccination is highly effective in reducing RSV-associated hospitalisations in older adults, consistent with findings from the pivotal clinical trial and emerging real-world studies. The evidence supports the introduction of RSV vaccination programmes for older populations in the UK where uptake can be optimised through targeted campaigns. Our findings highlight the importance of coordinated public health strategies, including clear communication to healthcare professionals and the public, to achieve high vaccine coverage and reduce hospital admissions. These data also provide a benchmark for monitoring vaccine impact and inform decisions regarding prioritisation, timing, and resource allocation in ongoing and future RSV immunisation programmes. Further real-world studies across diverse geographic and healthcare settings will be important to determine long-term effectiveness and inform policy.

Introduction

Although most commonly associated with infants, respiratory syncytial virus (RSV) also imposes a substantial burden on older adults. RSV can cause severe lower respiratory tract infection (LRTI) including pneumonia, particularly in individuals with underlying respiratory, cardiac, or immunocompromising conditions.1, 2, 3, 4 Recent UK data underscore the disproportionate burden of RSV in older adults, with seasonal averages of 70 respiratory hospital admissions per 100,000 adults aged 75–79 years (95% confidence interval [CI] 58–82) and 108 per 100,000 aged ≥75 years (95% CI 96–121).⁵ By comparison, influenza accounted for 128 hospital admissions per 100,000 adults aged 75–79 years (95% CI 112–146) and 203 per 100,000 adults aged ≥75 years (95% CI 185–222).⁵ RSV follows a well-defined seasonal pattern in the UK, with transmission typically occurring between October and February and peaking in December,⁶ consistent with other temperate regions.⁷

After decades of vaccine development,^8,9 2023 marked a milestone with the licensure of two RSV vaccines, the adjuvanted RSVpreF3 (Arexvy, GSK) and the bivalent RSVpreF (Abrysvo®, Pfizer). Both demonstrated high efficacy and safety in large phase 3 randomised trials involving over 50,000 participants across multiple countries.10, 11, 12 At the end of one season, RSVpreF3 achieved 83% efficacy (95% CI 58–94) against laboratory-confirmed RSV lower respiratory tract infection (LRTI), and 94% (95% CI 62–99) against severe RSV-related LRTI.¹⁰ RSVpreF showed comparable efficacy of 86% (95% CI 32–99) against RSV LRTI with three or more symptoms in adults aged 60 and older.¹¹ These data supported regulatory approvals across Europe¹³ and in conjunction with transmission dynamic cost effectiveness analyses led to the UK's Joint Committee on Vaccination and Immunisation (JCVI) in 2023 advising the introduction of a national RSV immunisation programme for older adults.^14,15 The programme, adopted by all four UK nations (England,¹⁶ Wales,¹⁷ Scotland,¹⁸ and Northern Ireland (NI)¹⁹), offers a single dose of RSVpreF (Abrysvo®, Pfizer) free of charge through the National Health Service (NHS) to individuals turning 75 years, with a catch-up campaign for those already 75–79 years; JCVI having noted limited trial data in 80 year olds and above. Eligibility criteria were broadly consistent across the four nations; however, as health policy and practical implementation are different in the nations, there were minor variations in how the programme was implemented, including differences in start dates and delivery arrangements.16, 17, 18, 19 In England, Wales, and NI, vaccination was offered to individuals on or after their 75th birthday, with eligibility defined as those aged 75–79 years on September 1, 2024 or turning 75 thereafter. In Scotland, eligibility was defined as adults aged 74–79 years on August, 1 2024, with individuals who would turn 75 at any point between August 1, 2024, and July, 31 2025 eligible immediately, without waiting until their 75th birthday. As of May 2025, uptake of the RSV vaccine among eligible adults was 58.9% in England,²⁰ 44.9% in Wales,²⁰ 70.6% in Scotland,²⁰ and 51.3% in NI.²¹

Although the RSVpreF vaccine demonstrated significant efficacy in preventing RSV-associated LRTI and acute respiratory illness among older adults,^11,12 important gaps remain regarding real-world vaccine effectiveness (VE) across diverse populations and healthcare settings. Post-licensure studies to date have been geographically restricted to the USA and largely confined to selected populations such as veterans or integrated health systems.22, 23, 24, 25, 26 Our study expands this evidence base by evaluating RSV VE across four UK nations using national surveillance data. The coordinated RSV vaccination programmes in England, Wales, Scotland, and NI provide a unique opportunity to generate robust, population-level evidence across different regions and healthcare systems to inform vaccination policy. This collaborative approach builds on successful precedent set by previous UK-wide studies leveraging pooled national data to produce high-quality epidemiological insights to inform public health policy, and provide robust evidence on vaccine performance and programme implementation.^27,28 The aim of this study is to estimate the real-world effectiveness of the RSVpreF vaccine in older adults years by conducting a meta-analysis of national test-negative design (TND) studies across the four UK nations.

Methods

Study design and population

We conducted a retrospective observational study using a test-negative case–control design to estimate the effectiveness of the RSVpreF vaccine in preventing hospitalisation due to laboratory-confirmed RSV infection among older adults. All TND analyses were conducted separately within each UK nation by the responsible public health agencies: the UK Health Security Agency (UKHSA) in England, Public Health Wales (PHW), Public Health Scotland (PHS), and the Public Health Agency (PHA) in NI, following a common protocol (Appendix pp 3–14). We combined the results of four nation-specific analyses instead of one single analysis because each nation's public health authority is responsible for its own data, and sharing individual-level data was not possible. Each nation used routinely collected national surveillance health datasets to identify eligible cases and controls. Results from nation-specific analyses were then combined in a UK-wide fixed-effects meta-analysis.

The study population consisted of individuals aged 74–79 years and eligible for preF NHS vaccination, who were hospitalised and tested for RSV using molecular methods. Implementation of the vaccination programme was undertaken separately by England, Wales, Scotland, and NI, resulting in slightly different eligibility criteria between nations. In England, Wales, and NI, inclusion was restricted to individuals aged 75–79 years on 1 September 2024 or turning 75 thereafter, with vaccination eligibility beginning on or after their 75th birthday. In Scotland, eligibility extended to adults aged 74–79 years on August, 1, 2024, with those turning 75 between August 1, 2024, and July 31, 2025, eligible immediately, without waiting until their 75th birthday. Inclusion criteria required a valid national health identifier—the NHS number in England and Wales, the Community Health Index (CHI) in Scotland, and the Health & Care Number (HCN) in NI—to enable dataset linkage. Swabs had to be collected within 14 days prior to admission or within two days after admission, and testing had to be reported through national surveillance systems. Individuals were excluded if the swab sample date was unknown or outside the study period; if vaccination occurred within 0–13 days before swab collection; if vaccination was recorded prior to the commencement of the RSV vaccination programme; or (for the primary analysis) if controls tested positive for SARS-CoV-2 or influenza.

Cases were defined as individuals meeting the age, vaccination eligibility, and admission criteria who had a respiratory swab with a positive test for RSV within the specified time window around admission. Controls were individuals meeting the same criteria who tested negative for RSV, with those testing positive for SARS-CoV-2 or influenza excluded. Each participant contributed only one RSV test linked to a single hospitalisation episode. For individuals with multiple tests, the first positive RSV result was prioritised. For each individual control without any positive test, the first negative test from a new hospital admission was retained, and one test was randomly sampled. Voided or duplicate tests were excluded.

Study period

To ensure methodological consistency across all four nations, the study period was defined as spanning epidemiological week 40 of 2024 (beginning 30 September 2024) through epidemiological week 13 of 2025 (ending 30 March 2025). This period captures the full RSV seasonal activity and aligns broadly with the national RSV vaccine rollout for older adults. Although the RSV vaccine rollout commenced on 1 September in England, NI, and Wales, and earlier on 1 August 2024 in Scotland, standardising the study period across all four nations facilitated harmonised comparative analyses.

Data sources

Full data sources are described in the Appendix p 15. In brief, all four public health authorities used routine surveillance laboratory, vaccine and hospital admissions datasets to derive a study dataset for their respective nation.

Intervention

The intervention under evaluation was a single dose of Pfizer's Abrysvo® (RSVpreF), a recombinant prefusion F (preF) protein-based vaccine containing antigens from both RSV subtypes A and B. Implementation of the vaccination programme is undertaken separately by England, NI, Scotland and Wales, resulting in slightly different eligibility criteria between nations: in England, Wales, and NI, eligible adults were those aged 75–79 on 1 September 2024 or turning 75 thereafter; in Scotland, eligibility was defined as those aged 74–79 on 1 August 2024 i.e. includes those turning 75 within the 12 months from programme launch and on a rolling basis.

Exposure

The exposure of interest was receipt of the RSVpreF vaccine administered as part of the national adult RSV immunisation programme. Vaccination status and dates were obtained from national immunisation registries in each nation: Immunisation Information System (IIS) in England, Vaccination Management Tool (VMT) in Scotland, Welsh Immunisation System (WIS) in Wales and Vaccine Management System (VMS) in NI (Appendix p 15). Vaccination status was categorised based on the interval between vaccine administration and respiratory swabbing. Individuals were considered vaccinated if they received the RSVpreF vaccine at least 14 days prior to respiratory sample collection. Vaccination within 0–13 days before sample collection was excluded due to uncertain immunogenicity in this interval. Vaccine exposure was linked at the individual level to testing and hospitalisation records using unique national health identifiers.

Outcome

The primary outcome was hospitalisation with laboratory-confirmed RSV infection. RSV infection was confirmed by molecular testing of respiratory samples collected within 14 days before hospital admission or up to two days after admission, as reported through national surveillance systems. Molecular methods primarily included reverse-transcription polymerase chain reaction (RT-PCR) assays. RSV-related hospitalisations were defined differently according to data availability. In England and Scotland, cases were restricted to those admitted with a LRTI diagnosis, identified through specific discharge codes (Appendix p 5). In NI and Wales, where hospital coding data were not promptly available, cases were defined as hospital admissions with RSV detected within the same 14-day pre-admission to two-day post-admission window, without reference to discharge.

Statistical analysis

Multivariable logistic regression models were used to compare the odds of RSV vaccination between case and control patients. Primary analyses in each nation were adjusted for epidemiological week of specimen collection using spline functions to account for seasonal variation in RSV circulation. Where data were available, additional nation-specific models were adjusted for individual-level covariates: Northern Ireland and Wales for age and sex; England additionally for care home status, clinical risk group, and deprivation; and Scotland for age, sex, clinical risk group, and deprivation. Data on race and ethnicity were not consistently available across the national surveillance datasets used in England, Wales, Scotland, and NI and were therefore not included in the analyses.

VE was calculated as (1—odds ratio) × 100%, with 95% confidence intervals (CIs) derived using the Wald method. Statistical significance was defined as a two-sided p-value <0.05. Nation-specific VE estimates were combined using fixed-effects meta-analysis with inverse-variance weighting,²⁹ with fixed-effects selected (as specified in the protocol) following heterogeneity assessment using the I² statistic. All analyses were performed using R and STATA.

Sensitivity analyses

Three sensitivity analyses were undertaken, two of which redefined the control group. In the first, individuals who tested positive for SARS-CoV-2 and influenza were included as controls to evaluate whether their inclusion bias the VE estimates, as suggested in TND studies due to correlated vaccination behaviour.³⁰ Second, influenza and SARS-CoV-2-positive controls were again included, but with additional adjustment for influenza and COVID-19 vaccination status. This approach accounts for potential confounding arising from correlated vaccination behaviours, which can bias VE estimates in test-negative studies when controls include vaccine-preventable respiratory infections.³⁰ Finally, to assess whether results were disproportionately influenced by data from any single UK nation, we conducted a leave-one-nation-out sensitivity analysis.³¹ For each iteration, we re-ran the fixed-effects meta-analysis excluding one nation at a time and recalculated the pooled VE. Although between-nation heterogeneity was negligible (I² = 0), this analysis was undertaken as a robustness check, recognising that larger populations (e.g., England) could still exert greater influence on the overall estimate through weighting.

Ethics

This work did not require research ethics approval because it was defined as Health Surveillance, not research, under the Health Research Authority's guidance (https://www.hra-decisiontools.org.uk/research/docs/DefiningResearchTable_Oct2022.pdf) and decision support tool. The work was conducted as part of routine infectious disease and vaccine surveillance programmes as a public task function of public health authorities to evaluate the vaccine programmes in the populations for which they are responsible, under their organisational governance arrangements and their legal bases for processing administrative personal healthcare data for health surveillance without consent. Legal bases and governance arrangements are specific to each UK country and organisation. In England, the work was carried out under Regulation 3 of The Health Service (Control of Patient Information) Regulations; Secretary of State for Health, 2002) using patient identification information without individual patient consent as part of the UKHSA mandate for public health surveillance and monitoring of vaccines. Public Health Wales NHS Trust, used patient identification information without individual patient consent, in line with the core function in article 3(2a) of the trust's establishment order. In Scotland, access to primary care data is covered under the National Health Service (Scotland) Act 1978 and the Public Health Data (Infectious Respiratory Diseases) (Scotland) Directions 2024.

Access is restricted to Public Health Scotland and is governed through established information governance frameworks and data protection agreements. Public Health Scotland and the Public Health Agency in Northern Ireland processed pseudonymised routine administrative healthcare data without patient consent under information governance agreements with data controllers as part of their statutory public task functions.

Reporting

This study is reported in accordance with the REporting of studies Conducted using Observational Routinely-collected Data guidelines (www.RECORD-statement.org; Appendix pp 16–21).

Funding

This study received no specific funding.

Results

The study population (n = 11,117) included 6528 individuals in England, 942 in NI, 1462 in Scotland, and 2185 in Wales who met the inclusion criteria during the study period. Table 1 summarises the characteristics of the study population and cases by nation. The number of unvaccinated individuals was 4245 (65.0%) in England, 598 (63.5%) in NI, 605 (41.4%) in Scotland, 1773 (81.1%) in Wales, and 7221 (65.0%) across the combined study population. Of the 826 cases across all nations, 726 (87.9%) were unvaccinated, including 607 (87.7%) in England, 20 (87.0%) in NI, 31 (77.5%) in Scotland, and 68 (95.8%) in Wales. Weekly numbers of RSV-positive cases and controls are shown in the Appendix (p 22).

Table 1.

Characteristics of the study population, by nation.

Characteristic	All nations		England		Northern Ireland		Scotland		Wales
	Patients, no (%)		Patients, no (%)		Patients, no (%)		Patients, no (%)		Patients, no (%)
	Total	Cases	Total	Cases	Total	Cases	Total	Cases	Total	Cases
Age
74	233 (2.1)	10 (1.2)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	233 (15.9)	10 (25.0)	0 (0.0)	0 (0.0)
75	2031 (18.3)	133 (16.1)	1198 (18.4)	116 (16.8)	191 (20.3)	2 (8.7)	239 (16.3)	4 (10.0)	403 (18.4)	11 (15.5)
76	2146 (19.3)	168 (20.3)	1271 (19.5)	143 (20.7)	195 (20.7)	6 (26.1)	219 (15.0)	5 (12.5)	461 (21.1)	14 (19.7)
77	2452 (22.1)	172 (20.8)	1486 (22.8)	143 (20.7)	199 (21.1)	7 (30.4)	294 (20.1)	5 (12.5)	473 (21.6)	17 (23.9)
78	2187 (19.7)	191 (23.1)	1307 (20.0)	160 (23.1)	177 (18.8)	4 (17.4)	259 (17.7)	10 (25.0)	444 (20.3)	17 (23.9)
79	2068 (18.6)	152 (18.4)	1266 (19.4)	130 (18.8)	180 (19.1)	4 (17.4)	218 (14.9)	6 (15.0)	404 (18.5)	12 (16.9)
Sex
Male	5650 (50.8)	358 (43.3)	3308 (50.7)	312 (45.1)	502 (53.3)	7 (30.4)	758 (51.8)	17 (42.5)	1082 (49.5)	22 (31)
Female	5467 (49.2)	468 (56.7)	3220 (49.3)	380 (54.9)	440 (46.7)	16 (69.6)	704 (48.2)	23 (57.5)	1103 (50.5)	49 (69)
Month of swab
September	247 (2.2)	3 (0.4)	12 (0.2)	1 (0.1)	9 (1.0)	0 (0.0)	0 (0.0)	0 (0.0)	226 (10.3)	2 (2.8)
October	1309 (11.8)	40 (4.8)	581 (8.9)	33 (4.8)	149 (15.8)	1 (4.3)	255 (17.4)	0 (0.0)	324 (14.8)	6 (8.5)
November	1715 (15.4)	182 (22.0)	1058 (16.2)	158 (22.8)	153 (16.2)	3 (13.0)	227 (15.5)	5 (12.5)	277 (12.7)	16 (22.5)
December	2332 (21.0)	315 (38.1)	1497 (22.9)	269 (38.9)	183 (19.4)	11 (47.8)	317 (21.7)	13 (32.5)	335 (15.3)	22 (31.0)
January	2365 (21.3)	191 (23.1)	1463 (22.4)	151 (21.8)	168 (17.8)	6 (26.1)	306 (20.9)	18 (45.0)	428 (19.6)	16 (22.5)
February	1574 (14.2)	66 (8.0)	934 (14.3)	55 (7.9)	144 (15.3)	1 (4.3)	190 (13.0)	3 (7.5)	306 (14.0)	7 (9.9)
March	1575 (14.2)	29 (3.5)	983 (15.1)	25 (3.6)	136 (14.4)	1 (4.3)	167 (11.4)	1 (2.5)	289 (13.2)	2 (2.8)
Vaccination status
Vaccinated	3896 (35.0)	100 (12.1)	2283 (35.0)	85 (12.3)	344 (36.5)	3 (13.0)	857 (58.6)	9 (22.5)	412 (18.9)	3 (4.2)
Not vaccinated	7221 (65.0)	726 (87.9)	4245 (65.0)	607 (87.7)	598 (63.5)	20 (87.0)	605 (41.4)	31 (77.5)	1773 (81.1)	68 (95.8)

Data are n (%). Percentages are calculated within each stratum. Characteristics shown include age, sex, month of swab, and vaccination status. Vaccination status refers to receipt of a single dose of RSVpreF (Abrysvo®). In England, Wales, and Northern Ireland, vaccination was offered on or after an individual's 75th birthday, and only adults aged 75–79 years are included. In Scotland, adults aged 74–79 years were eligible, with those turning 75 between August 2024 and July 2026 eligible immediately, without waiting until their 75th birthday.

Across the four nations, RSVpreF vaccination was associated with a significant reduction in the odds of RSV-related hospitalisation among older adults. Primary analyses, adjusted for epidemiological week of specimen collection, showed VE estimates of 74% (68–80) in England, 77% (30–95) in Northern Ireland, 81% (60–91) in Scotland, and 82% (42–94) in Wales, with a pooled fixed-effects estimate of 75% (69–80) (Fig. 1). Additional meta-analyses adjusted for individual-level covariates could not be undertaken because relevant data were not consistently available across all four nations. Instead, nation-specific analyses were performed using available covariates: NI and Wales adjusted for age and sex; England additionally for care home status, clinical risk group, and deprivation; and Scotland for age, sex, clinical risk group, and deprivation. These adjustments had little impact on VE estimates (Appendix p 23). In England, where immunosuppression status was available, VE was 65% (48–77) in immunocompromised individuals and 78% (71–84) in those without immunosuppression.

Fig. 1.

Bivalent RSV prefusion vaccine effectiveness (VE) against RSV-related hospitalisation, individual nation estimates, and meta-analysis. VE estimated i) excluding SARS-CoV-2 and influenza-positive individuals as controls (Main analysis) ii) including SARS-CoV-2 and influenza-positive controls with no adjustment for influenza and COVID-19 vaccination status (Sensitivity analysis 1), and iii) including SARS-CoV-2 and influenza-positive controls with additional adjustment for influenza and COVID-19 vaccination status (Sensitivity analysis 2). Pooled estimates were calculated using fixed-effects meta-analysis with inverse-variance weighting.

To address potential differences between nations, including differences in outcome definitions, we conducted a sensitivity analysis using a random-effects meta-analysis. The pooled estimate was unchanged, with overlapping confidence intervals, supporting the primary fixed-effects model, which was therefore retained given the absence of measurable between-nation heterogeneity. Additional sensitivity analyses, all of which also adjusted for epidemiological week, assessed whether the choice of control group or correlated vaccination behaviours influenced VE estimates. Wales was excluded from these analyses because influenza and COVID-19 vaccination status was not available in their linked datasets. Including SARS-CoV-2 and influenza-positive individuals as controls produced a pooled VE of 74% (68–80), while further adjustment for influenza and COVID-19 vaccination status yielded a pooled VE of 77% (71–82). VE estimates were consistent with the primary analysis, and heterogeneity across nations was low in all models (I² = 0.0%) (Fig. 1). These findings indicate that VE estimates are robust to alternative control definitions and adjustment for correlated vaccination behaviours.

In the leave-one-nation-out sensitivity analysis, pooled VE estimates were recalculated iteratively, excluding one nation at a time. Excluding England, which contributed the largest sample, increased the pooled VE to 81% (66–89), reflecting England's weighting in the overall estimate (Appendix p 24). Exclusion of NI, Scotland, or Wales had minimal impact, with pooled VE estimates remaining at 75% (95% CI 68–80 to 69–80) and I² = 0% in all iterations. VE from England was additionally calculated separately for the SGSS and DataMart systems and no difference found (73%, 95% CI: 66–79; 73%, 95% CI: 55–85 respectively). Heterogeneity across nations remained negligible in all sensitivity analyses, supporting the consistency of vaccine protection across the four UK nations.

Discussion

This study presents the first UK-wide evaluation of bivalent RSVpreF (Abrysvo®, Pfizer) VE against hospitalisation with laboratory-confirmed RSV infection in older adults during the 2024–25 season following introduction of the vaccine across all four UK nations. Leveraging a harmonised four-nation protocol and individual-level electronic health records covering the full eligible population, this analysis provides robust, population-level evidence beyond what has been possible in previous studies restricted to sentinel sites or selected populations. We found that RSV vaccination in adults aged 75–79 years was associated with a substantially reduced odds of RSV-associated hospitalisation, with a pooled VE of 75% (69–80). Results were broadly consistent across nations despite differences in healthcare systems.

Our findings align with pre-licensure trial results demonstrating strong protection offered by Pfizer RSVpreF vaccines.¹¹ Our pooled vaccine effectiveness of 75% (69–80) confirms that the high level of protection observed in trials is maintained in routine clinical practice. Given that approximately 70 per 100,000 RSV-related hospitalisations occur annually among UK adults aged 75–79 years (and over 108 per 100,000 ≥ 80 years),⁵ the potential population-level impact of vaccination is substantial. Early national impact evaluations provide supporting evidence: in England, introduction of the RSV vaccination programme was associated with a 30% reduction in hospitalisations among 75–79 year olds eligible for vaccination,³² and in Scotland, where uptake was higher, a 62% reduction in RSV-related hospitalisations was seen in the same age group following programme rollout in August, 2024.³³ These findings highlight the value of immunisation against RSV in reducing severe disease burden in older adults.

Several observational studies using TND have reported comparable VE estimates.22, 23, 24, 25 In the USA, Payne et al.²² found 80% (71–85) VE against RSV-associated hospitalisation in adults aged ≥60 years without immunocompromising conditions, and 73% (48–85) among those with immunocompromising conditions. Estimates from England found a similar difference, with overlapping confidence intervals around VE for each group. VE estimates were similar across age groups and vaccine products, with analyses adjusted for age, race and ethnicity, sex, calendar day, social vulnerability index, number of underlying non-respiratory medical conditions, presence of respiratory underlying medical conditions, and geographical region. Also, in the USA, Tartof et al.²³ evaluated RSVpreF effectiveness in adults aged ≥60 years using two prespecified control groups to ensure robustness: the strict control group included LRTI events that tested negative not only for RSV but also for human metapneumovirus, influenza, and SARS-CoV-2, and were positive for a nonvaccine-preventable disease. The broad control group included all RSV-negative LRTI events regardless of other identified causes. VE estimates were consistent between these two control groups, with effectiveness of 89% (95% CI 52–97) and 87% (95% CI 48–97) against RSV-related hospitalisations and emergency department visits, respectively, after adjusting for month of encounter, age, sex, self-reported race, ethnicity, comorbidity burden, and health care utilisation in the year before encounter. Surie et al.²⁴ estimated VE at 75% (50–87) in adults aged ≥60 years, with similar protection in those aged 60–74 years (75% (31–91)) and ≥75 years (76% (40–91)).

In England, data from the hospital-based acute respiratory infection sentinel surveillance (HARISS) system across 14 hospitals showed VE of 82% (71–90) against RSV-associated ARI hospitalisation in adults aged 75–79 years, and 89% (76–96) for LRTI hospitalisations.³⁴ HARISS sites are part of an active enhanced surveillance system chosen because of their systematic use of the respiratory PCR panel for COVID-19, influenza, and RSV for those presenting with respiratory symptoms, with prospective (admission) enrolment rather than retrospective from discharge coding, and a minimum length of stay requirement that may select for more severe disease. The HARISS study, however, represents only a small subset of hospitals in England: in the present study, 6528 adults admitted with an LRTI diagnosis were included from England, compared with 664 in the previously reported HARISS analysis. Moreover, England-specific HARISS findings do not capture the full UK context, given differences in health service organisation, vaccination delivery models, population structure, and data systems across the four nations. By applying a harmonised test-negative design with aligned covariate adjustment across England, Scotland, Wales, and NI, this study draws on a substantially larger underlying population and provides the first pooled UK-wide real-world estimate of RSVpreF vaccine effectiveness. This approach improves precision and representativeness, and establishes the baseline UK benchmark needed to assess waning immunity, revaccination strategies, and seasonal variation, generating evidence directly relevant to JCVI. It remains to be determined whether findings from TND studies with enhanced clinical surveillance and/or testing^5,23 differ systematically from those based on routine data,^22,24 including this work, noting confidence intervals substantially overlap. No difference was seen in England estimates between SGSS with newly activated negative reporting and the established DataMart sentinel laboratory network. These considerations underline the utility of pooled VE estimates in the context of diverse healthcare settings, reflecting the real-world impact of vaccination at a national level. Emerging evidence also indicates that protection following RSVpreF vaccination declines gradually over time: neutralising antibody titres peak around one month after vaccination and then fall over the subsequent year, although they remain above baseline; corresponding modest reductions in VE have been observed in early post-licensure analyses.³⁵ These findings emphasise the importance of continued multi-season monitoring to characterise waning, determine the durability of protection, and inform potential revaccination strategies for older adults.

A key strength of this study is its population-based, test-negative design with extensive coverage of the target population through multiple national surveillance programmes. Parallel analyses across four UK nations using a harmonised protocol minimised methodological variability. Nation-specific estimates were then combined using fixed-effects meta-analysis to provide pooled VE. This design supports interpretation of VE in a UK-wide policy context. Limitations include the potential for residual confounding from unmeasured factors, because relevant data were not available in a consistent format across all UK nations. As a result, we could not produce additional meta-analyses with uniform adjustment for confounders. However, nation-specific analyses were conducted: NI and Wales adjusted for age and sex; England adjusted for care home status, clinical risk group, and deprivation; and Scotland for age, sex, clinical risk group, and deprivation. Inclusion of these covariates had little impact on VE estimates, with adjusted results presented in the Appendix. Hospitalisation definitions were not fully uniform: England and Scotland restricted to LRTI-coded hospitalisations, whereas NI and Wales relied solely on laboratory confirmation of RSV by PCR within 14 days before to 2 days after admission due to unavailability of timely coding data. However, we did not see higher VE where LRTI was used. Wales did not contribute to the sensitivity analysis including influenza and SARS-CoV-2 controls, but findings from the other nations suggested minimal impact of this adjustment. While leave-one-out sensitivity analysis showed England to be the most influential nation on VE, overall study heterogeneity was low, and small changes in vaccinated case numbers would readily influence other UK nations’ estimates (not shown): i.e. between-nation differences in this study are consistent with chance variation. Data limitations also precluded VE stratification by RSV subtype or UK-level assessment in immunocompromised populations. In addition, linkage to ICU and mortality data was not possible for all four nations, precluding assessment of VE against the most severe outcomes.

Our findings support continued RSVpreF vaccination for older adults, with the potential to substantially reduce RSV-related hospitalisations and associated healthcare burden. Using the pooled VE of 75% and nation-specific uptake,^20,21 we estimate that approximately 548 admissions were averted in England, 36 in Wales, 45 in Scotland, and 14 in NI during the 2024–25 season. If uptake across the UK matched the higher level observed in Scotland, these figures would increase to around 779, 80, and 25 admissions in England, Wales, and NI, respectively. These projections highlight the substantial additional benefit achievable through improved vaccine coverage. Our study also establishes a UK-wide baseline from which durability of protection can be assessed. Although we could not reliably estimate VE by time since vaccination—owing to small numbers of RSV-related admissions in some nations—we plan follow-up analyses at the end of the 2025–26 season, when two years of post-implementation data will support evaluation of waning immunity and inform future revaccination strategies. In July 2025, the UK JCVI extended its advice for a single dose of RSV vaccine to those aged ≥80 years and residents of care homes for older adults.³⁶ Our results provide timely evidence to inform implementation of this policy.

Contributors

MBu: conceptualisation, methodology, investigation, formal analysis, visualisation, validation, writing—original draft, writing—review & editing; SSH: conceptualisation, methodology, data curation, visualization, formal analysis, writing—review & editing; AAM: conceptualisation, methodology, data curation, formal analysis, writing—review & editing; SC: conceptualisation, methodology, investigation, formal analysis, supervision, writing—review & editing; HJW: conceptualisation, methodology, data curation, formal analysis, validation, writing—review & editing; JBH: data curation, writing—review & editing; AC: data curation, writing—review & editing; NA: methodology, supervision, writing—review & editing; CHW: conceptualisation, methodology, data curation, supervision, writing—review & editing; DTB: methodology, supervision, writing—review & editing; MOD: data curation, writing—review & editing; ED: data curation, writing—review & editing; CB: data curation, writing—review & editing; JM: data curation, writing—review & editing; CJW: writing—review & editing; CM: writing—review & editing; MP: data curation, writing—review & editing; JZ: data curation, writing—review & editing; PK: data curation, writing—review & editing; MBa: data curation, writing—review & editing; CJ: data curation, writing—review & editing; PR: data curation, writing—review & editing; VM: data curation, writing—review & editing; KMo: conceptualisation, validation, writing—original draft, writing—review & editing; CR: conceptualisation, formal analysis, methodology, writing—review & editing; RMcQ: conceptualisation, writing—review & editing; KMa: conceptualisation, writing—review & editing; SG: conceptualisation, writing—review & editing; JMcM: conceptualisation, writing—review & editing. In each organisation, at least two authors accessed and verified the data. PHA, Northern Ireland: MBu, MoD. PHS, Scotland: SSH, KM, CR; UKHSA, England: AAM, HW. PHW, Wales, SC, MB. All authors revised the manuscript and approved the final version of the manuscript in advance of submission. All authors had full access to all the data within the public health authority for their own participating nation, and all authors had access to the aggregate results data that were combined by meta-analysis. All authors had final responsibility for the decision to submit for publication.

Data sharing statement

England: This work is carried out under Regulation 3 of The Health Service (Control of Patient Information) Regulations; Secretary of State for Health, 2002) using patient identification information without individual patient consent as part of the UKHSA legal requirement for public health surveillance and monitoring of vaccines. As such, authors cannot make the underlying dataset publicly available for ethical and legal reasons. However, all the data used for this analysis is included as aggregated data in the manuscript Table 1. Applications for relevant anonymised data should be submitted to the UKHSA Office for Data Release at https://www.gov.uk/government/publications/accessing-ukhsa-protected-data.

Northern Ireland: Health and Social Care data are available for use by approved researchers and internal HSC analysts by application to the Business Services Organisation Honest Broker Service Trusted Research Environment. This work was conducted on pseudonymised data as part of routine infectious disease surveillance under information governance agreements with data controllers.

Scotland: This work was carried out in Scotland on pseudonymised data as part of routine infectious disease surveillance by members of the Vaccine Effectiveness team in Public Health Scotland. Due to the terms of data access, individual-level data used in this analysis cannot be shared publicly.

Wales: This work was carried out in Wales by the specialists in the Vaccine Preventable Disease Programme and Communicable Disease Surveillance Centre of the Public Health Wales NHS Trust, using patient identification information without individual patient consent, in line with the core function in article 3(2a) of the trust's establishment order (to provide to or in relation to the health service in Wales and manage a range of public health, health protection, healthcare improvement, health advisory, child protection and microbiological laboratory services and services relating to the surveillance, prevention and control of communicable diseases). All data are held on NHS Wales secure servers and processed in accordance to Public Health Wales Information Governance policies.

Declaration of interests

CHW and NA report that the Immunisation and Vaccine Preventable Diseases Department at UKHSA has provided vaccine manufacturers, including Pfizer UK, with post-marketing surveillance reports, which marketing authorisation holders are required to submit to the UK licencing authority in compliance with their risk management strategies; a cost-recovery charge is made for these reports. CHW also reports support for conference grants from the British Thoracic Society and the UK Clinical Vaccines Network. CR reports a research grant to the University of Strathclyde and participation on a Data Safety Monitoring Board or Advisory Board for the International Breast Cancer Study Group (DSMC and Ethics), with personal honorarium. CR also reports membership of SPI-M.