Summary
Background
Seasonal influenza causes substantial morbidity and mortality in older adults. While vaccination is recommended in Nordic countries for individuals aged ≥65 years, brand-specific effectiveness estimates are scarce but essential for regulatory decision-making. We evaluated brand-specific effectiveness of seasonal influenza vaccines against laboratory-confirmed influenza-related outcomes in Denmark, Finland, and one Swedish region during the 2024/2025 season.
Methods
We conducted a nationwide cohort study using target trial emulation and linked health registries. Individuals aged ≥65 years were matched 1:1, comparing seasonal influenza vaccine recipients to non-recipients. Cumulative incidences of laboratory-confirmed influenza A and B, influenza-related hospitalisation, and death were assessed at 18 weeks post-immunisation. Vaccine effectiveness (VE) was calculated as 1 minus the risk ratio.
Findings
A total of 1,164,686 matched pairs were included (mean age 75.4 years, standard deviation 7.3). Overall VE against influenza-related hospitalisation was 46.8% (95% CI 40.8–52.9), with risk differences of −21.3 (−28.9 to −13.8) and −99.3 (−119.6 to −79.1) in Finland and Denmark, respectively, per 100,000 vaccinated individuals. Brand-specific VE was 63.4% (38.1–88.7) for Efluelda Tetra (split virion-high dose), 48.2% (40.8–55.6) for Fluad Tetra (subunit standard-dose adjuvanted), 43.6% (23.7–63.6) for Vaxigrip Tetra (split virion standard-dose), and 30.6% (−7.8 to 69.1) for Influvac Tetra (subunit standard-dose). VE waned by −6.5 percentage points (−10.5 to −2.5) every 3 weeks.
Interpretation
Seasonal influenza vaccines moderately reduced the risk of severe outcomes in older adults. Efluelda Tetra and Fluad Tetra appeared to offer favourable protection in their respective target groups, supporting their use in the 2025/2026 season. Annual monitoring using Nordic registries is crucial for informing evidence-based vaccination strategies and regulatory decisions.
Funding
European Medicines Agency.
Keywords: Influenza vaccines, Vaccine effectiveness, Target trial emulation, Registries
Research in context.
Evidence before this study
Seasonal influenza is a major public health concern, particularly for older adults, for whom annual vaccination is recommended in the Nordic countries. Although influenza vaccines are widely used, real-world, brand-specific effectiveness data remain limited. We searched PubMed for studies published between 2023 and 2025 using the terms (“influenza vaccines” OR “influenza vaccine”) AND (“vaccine efficacy” OR “effectiveness”), without language restrictions. Of 679 records screened, few reported brand-specific vaccine effectiveness (VE) estimates in adults aged 65 years and older.
Recent studies include an interim 2024/25 European analysis across 17 countries reporting VE of 41–62% against influenza subtypes in older adults using a test-negative design, and a 2023 Southern Hemisphere study across eight countries estimating VE of 47.7% against influenza-related hospitalisation. Two VEBIS multicentre studies from the 2022/23 and 2023/24 seasons provided subtype-specific VE across up to 12 European countries. A systematic review covering 53 million individuals and the DRIVE study across seven countries during 2021/22 highlighted the value of adjuvanted and high-dose vaccines and the feasibility of brand-specific VE estimates.
Added value of this study
During the 2024/25 influenza season in two Nordic countries, vaccination was associated with reduced risk of laboratory-confirmed influenza A and B, influenza-related hospitalisation, and death among adults aged 65 years and older. Using routinely collected nationwide registry data, this study provides timely, brand-specific VE estimates. Adjuvanted and high-dose vaccines appeared more effective than standard-dose quadrivalent vaccines. By combining target trial emulation and test-negative design approaches, the study demonstrates the feasibility of generating robust, brand-specific VE estimates through population-based linkable data systems. Supported by the European Medicines Agency, it also illustrates the regulatory and public health value of such evidence.
Implications of all the available evidence
This study supports the moderate effectiveness of seasonal influenza vaccination in older adults, with enhanced vaccines (adjuvanted and high-dose formulations) appearing to offer greater protection than standard-dose vaccines. These findings support prioritisation of adjuvanted and high-dose vaccines for adults aged 65 years and older, where cost-effective. However, head-to-head comparisons of brand-specific vaccine effectiveness could yield more precise estimates of their relative performance and such studies are warranted. Continued annual monitoring through linked health registries across multiple countries is essential for guiding optimal vaccine use, supporting product authorisation, and contributing to pandemic preparedness efforts.
Introduction
Seasonal influenza remains a major public health burden, particularly among older adults at increased risk of severe illness, hospitalisation, and death.1 During the 2024/25 season, Denmark alone reported over 12,000 laboratory-confirmed influenza cases, with most hospitalisations and deaths among people aged 65 years and older.2
Vaccination is the primary prevention strategy, and enhanced vaccines—such as adjuvanted and high-dose formulations—are now available alongside standard-dose products in the Nordic countries, with recommendations varying by age group and product. Despite the ongoing use of these vaccines in older adults, there is a lack of brand-specific, real-world VE estimates from northern Europe to inform programme decisions.
Previous studies in older adults in similar northern European settings have reported moderate protection against influenza outcomes. In a Finnish-Swedish study during the 2016/17 season, VE against laboratory-confirmed influenza was 24% (95% CI: 11–35%) in Stockholm and 33% (95% CI: 28–38%) in Finland.3 A Finnish register-based analysis during the 2017/18 season reported a VE of 46% (95% CI: 34–56%) against influenza A-related hospitalisation in adults ≥65 years.4 More recently, a 2021/22 cohort study within the multi-country DRIVE platform estimated VE for Vaxigrip Tetra at 81% (95% CI: 22–95%) against laboratory-confirmed influenza in Finnish adults aged ≥65 years. The study also found evidence of variation in effectiveness between standard-dose and enhanced products in this age group, demonstrating the feasibility and potential policy relevance of brand-specific VE estimation.5 Together, these studies highlight the importance of brand-specific evidence in shaping vaccination strategies, including consideration of preferential recommendations.
For the current 2024/25 season, interim results from eight European countries, including Denmark, reported VE of 49% (95% CI: 39–58%) against influenza A(H1N1) pdm09, 41% (95% CI: 30–50%) against A(H3N2), and 62% (95% CI: 46–73%) against influenza B in adults seeking primary care.6 While these estimates were not brand-specific, they provide context for interpreting the present study. Test-negative design (TND) studies dominate influenza VE research despite their limitations. Our study applies a target trial emulation (TTE) framework, providing a complementary perspective to TND-based estimates.
This study aims to estimate the brand-specific effectiveness of seasonal influenza vaccines in preventing laboratory-confirmed influenza outcomes among individuals aged 65 years and older in Denmark, Finland, and Sweden (1 region) during the 2024/25 season.
Methods
Data sources, study design, and cohort specification
In the three countries, we linked demography and healthcare data across different nationwide registries by using the country-specific unique identifiers assigned to all residents. We retrieved individual-level information on brand-specific influenza vaccinations, hospital admissions and recorded disease diagnoses, laboratory-confirmed influenza infections, and demographic (age, sex, and vital status) variables (Supplementary Tables S1 and S2). Sweden has no national registration of administered influenza vaccinations and only regional data from Uppsala Region (405,000 inhabitants) was available. Follow-up in Sweden was limited to 12 weeks (day 84 of follow-up) due to unavailability of data, and results are presented separately as a Supplementary Material.
We designed this non-interventional cohort study utilising the target trial emulation (TTE) framework (Supplementary Table S3).7 We compared the cumulative incidences of a positive polymerase chain reaction (PCR)-confirmed influenza test (influenza types A and B), hospital admission and death related to influenza, according to receiving or not receiving the seasonal influenza vaccination during the study period 1 October 2024 to 21 March 2025. Across the three Nordic countries, the influenza season started around weeks 49–50 of 2024, with influenza A predominating the study period; the 2024/2025 influenza season wave peaked between weeks 6 and 9 in 2025. In Denmark, vaccines were administered from October 1, in Finland from September 30 and in Sweden from October 15, until December 20, 2024, in all three countries. Quadrivalent vaccines administered this season comprised Vaxigrip Tetra (split virion standard-dose, SV-SD), Influvac Tetra (subunit standard-dose, SU-SD), Fluad Tetra (subunit standard-dose adjuvanted, SU-SD-Adj), and Efluelda Tetra (split virion high-dose, SV-HD) (Supplementary Table S4).
The following eligibility criteria were assessed at the start of the study period: age ≥65 years, and having a country residency (to ensure a linkable identifier); to construct a cohort representative of the general population targeted for vaccination with the seasonal influenza vaccine during autumn and winter 2024/2025 as per the national influenza vaccination strategies.
Outcomes
We defined an influenza A or B case as a laboratory-confirmed influenza infection, based on a positive PCR test for subtype A or B, respectively. Hospitalisation due to influenza was defined as a) hospitalisation with laboratory-confirmation within 14 days before to 2 days after the admission date, b) an inpatient contact or at least 12 h of contact, and c) influenza-like illness relevant diagnosis code (ICD-10: J09, J10, J11). Influenza-related death was defined as the date of death within 30 days after laboratory-confirmation of influenza. Individuals were excluded if an event occurred 90 days before the study start date (October 1, 2024) (Supplementary Table S5).
Procedures
Individuals who received the vaccine were matched on the day of vaccination with individuals who had not (yet) received the vaccine. Individuals were matched 1:1 on age (5-year bins), sex, region of residence, and number (0, 1, 2, or ≥3) of selected comorbidities (i.e., chronic pulmonary disease, cardiovascular conditions, diabetes, autoimmunity-related conditions, cancer, and moderate-to-severe renal disease) (Supplementary Table S2). The day the seasonal vaccine dose was administered within each matched pair served as the index date for both individuals, which inherently controlled for calendar time at study entry. We followed individuals from day 14 after the index date (to ensure full immunisation) up until the day of an outcome event, death, emigration, or end of the study, whichever occurred first for each individual. We censored individuals with a positive PCR test for influenza in our follow-up period 14 and 30 days after the test in the influenza hospitalisation and death outcome analyses, respectively.
Statistical analysis
We used the Aalen-Johansen estimator to obtain cumulative incidences of the outcomes among seasonal influenza vaccine recipients and non-recipients. Relative (vaccine effectiveness; calculated as 1–risk ratio) and absolute (the estimated number of outcome events prevented by vaccination; reported per 100,000 vaccinated individuals) risk differences (RD) were calculated from the cumulative incidences at 18 weeks (day 126 of follow-up), determined by data availability. The corresponding 95% confidence intervals (CI) were calculated using the delta method, and truncated if higher than 100%. We combined country-specific VE estimates by random-effects meta-analyses implemented using the mixmeta package in R using the restricted maximum likelihood estimation method.8
Subgroup analyses were done by sex (female/male), age groups (65–74/≥75 years), and vaccine brand. Changes in VE during follow-up (i.e., waning vaccine-induced immunity) were assessed by stratifying follow-up in three-week intervals for both recipients and non-recipients to obtain VE estimates for every three-week period separately. These estimates were analysed with linear regression, where the slope coefficient represented the per three-week percentage point change in VE.9
To complement our main analysis using target trial emulation, we conducted a supplementary test-negative design (TND) analysis to estimate VE against laboratory-confirmed influenza A using Danish data. Negative tests within 7 days of a previous negative test and within 21 days before a subsequent positive test were excluded, as were all tests within 90 days of a previous positive test, to avoid counting the same episode twice. Individuals were considered vaccinated if they had received an influenza vaccine at least 14 days before the PCR sample date, and unvaccinated if they had not received the vaccine or had been vaccinated within the preceding 14 days. Individuals with a PCR sample date between their vaccination date and 14 days thereafter, or more than 18 weeks after vaccination, were excluded. Vaccine effectiveness was estimated as 1 − odds ratio (OR), with ORs obtained from multivariable logistic regression models adjusted for age, sex, region of residence, comorbidities, and week of test.
To evaluate the possible residual confounding in the main analysis, we analysed the association between influenza vaccination and the following possible negative control outcomes (NCOs) in our main study design: lower back pain (ICD10: M543-M545), clavicle fracture (ICD10: S420), diverticulitis (ICD: K57), and all-cause mortality, defined here as death from any cause among any participant in the cohort.10 The results are presented as cumulative incidence curves with RD and VE estimates. For these outcomes to be true negative control outcomes, the associations between vaccination and the outcomes need to be subject to the same potential confounder(s) as the association between vaccination and the study outcome. E-values using the R package EValue were calculated to quantify the potential bias needed to fully explain the main and negative control outcome estimates.
Due to completeness of nationwide registries, there were no missing data on the variables included in this study. This study is reported in accordance with the STROBE guidelines.11
Ethical statement
The study was conducted in accordance with national regulations in Denmark, Finland, and Sweden. In Denmark, the analyses were performed as part of statutory surveillance activities by Statens Serum Institut and did not require ethics committee approval. In Finland, the Finnish Institute for Health and Welfare is mandated by law to conduct vaccination surveillance, and a waiver of ethical approval was granted by the Director of the Department for Health Security. In Sweden, the study was approved by the Swedish Ethical Review Authority (Dnr 2020-06859 and 2021-02186) and conducted in line with the Declaration of Helsinki; register-based studies are exempt from informed consent requirements.
Role of the funding source
This research was supported by the European Medicines Agency. The funders had no role in deciding the study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the article for publication. This document expresses the opinion of the authors of the paper and may not be understood or quoted as being made on behalf of, or reflecting the position of, the European Medicines Agency or one of its committees or working parties.
Results
Study population
Supplementary Figure S1 outlines the selection of cohort participants. Prior to matching, the source cohorts comprised 1,611,962 recipients of the seasonal influenza vaccine in the three countries during the study period. The largest number of recipients were from Denmark (826,766), followed by Finland (752,350) (Table 1).
Table 1.
Cohort characteristics before and 14 days after matching of seasonal influenza vaccine recipients and non-recipients aged ≥65 years in Denmark, Finland, and Sweden, 1 October 2024 to 21 March 2025.a, d
| Before matching |
After matching |
|||
|---|---|---|---|---|
| Seasonal recipient | Non-recipientb | Seasonal recipient | Non-recipientb | |
| No. individuals | ||||
| Total | 1,611,962 | 2,520,189 | 1,164,686 | 1,164,686 |
| Denmark | 826,766 | 1,129,872 | 529,082 | 529,082 |
| Finland | 752,350 | 1,311,727 | 611,174 | 611,174 |
| Sweden | 32,846 | 78,590 | 24,430 | 24,430 |
| Demographics | ||||
| Mean age (SD), years | 75.9 (7.3) | 75.3 (7.4) | 75.4 (7.3) | 75.4 (7.3) |
| Female sex | 890,814 (55.3) | 1,382,276 (54.8) | 647,756 (55.6) | 647,756 (55.6) |
| Comorbidities | ||||
| Autoimmune related conditions | 91,001 (5.6) | 129,219 (5.1) | 61,727 (5.3) | 58,011 (5.0) |
| Cancer | 216,473 (13.4) | 314,711 (12.5) | 150,410 (12.9) | 142,635 (12.2) |
| Chronic pulmonary disease | 86,937 (5.4) | 120,588 (4.8) | 56,345 (4.8) | 54,110 (4.6) |
| Cardiovascular condition | 441,175 (27.4) | 670,035 (26.6) | 311,783 (26.8) | 317,891 (27.3) |
| Diabetes | 163,374 (10.1) | 256,511 (10.2) | 117,826 (10.1) | 123,804 (10.6) |
| Renal disease | 46,608 (2.9) | 68,725 (2.7) | 30,427 (2.6) | 32,150 (2.8) |
| Comorbidity count | ||||
| 0 | 857,776 (53.2) | 1,389,682 (55.1) | 635,941 (54.6) | 635,941 (54.6) |
| 1 | 520,209 (32.3) | 784,994 (31.1) | 367,483 (31.6) | 367,483 (31.6) |
| 2 | 184,682 (11.5) | 273,596 (10.9) | 128,104 (11.0) | 128,104 (11.0) |
| >2 | 49,295 (3.1) | 71,917 (2.9) | 33,158 (2.8) | 33,158 (2.8) |
| Vaccine brands | ||||
| Influvac Tetra (SU-SD) | 44,795 (2.8) | 36,098 (3.1) | ||
| Vaxigrip Tetra (SV-SD) | 854,562 (53.0) | 688,822 (59.1) | ||
| Efluelda Tetra (SV-HD) | 51,871 (3.2) | 37,246 (3.2) | ||
| Fluad Tetra (SU-SD-Adj) | 659,153 (40.9) | 402,490 (34.6) | ||
| Flucelvax Tetrac | 40 (0.0) | 20 (0.0) | ||
| Fluarix Tetrac | 46 (0.0) | 0 (0.0) | ||
| Fluenz Tetrac | 3 (0.0) | |||
| Fluenzc | 15 (0.0) | 10 (0.0) | ||
| Fluzonec | 6 (0.0) | |||
Values are numbers (percentages) unless stated otherwise.
Individuals eligible for influenza vaccination as of the start of the study on 1 October 2024.
Analysis could not be performed due to a low number of individuals vaccinated and/or a low number of events in these subgroups.
Per study design, we applied a pairwise censoring if a reference individual received a vaccine. This led to 329,160 matched pairs experiencing censoring prior to start of follow up at 14 days after vaccination due to quick vaccine uptake.
The matched cohorts consisted of a total of 1,164,686 recipients during the study period (mean age 75.4, standard deviation 7.3 years) and 1,164,686 non-recipients. The most recipients were from Finland (611,174), followed by Denmark (529,082). The most frequently used vaccine brand was Vaxigrip Tetra (SV-SD) (688,822 doses, 59.1%), followed by Fluad Tetra (SU-SD-Adj) (402,490 doses, 34.6%), Efluelda Tetra (SV-HD) (37,246 doses, 3.2%), and Influvac Tetra (SU-SD) (36,098 doses, 3.1%) (Table 1). Characteristics among the matched pairs were overall similar to those of the unmatched populations (Supplementary Figure S2).
Brand-specific effectiveness of the seasonal influenza vaccine
Below we present VE and RD estimates from Denmark and Finland at week 18. The estimates at week 12 including results from Sweden, are presented in Supplementary Figure S3. I2 values to quantify the heterogeneity across studies can be found in Supplementary Table S11.
Laboratory-confirmed influenza A
Vaccine recipients had lower cumulative incidence of laboratory-confirmed influenza A compared to non-recipients (Fig. 1).
Fig. 1.
Cumulative incidence curves of laboratory-confirmed influenza A and B, influenza hospitalisation, and influenza-related death comparing recipients of influenza vaccine during 2024–2025 season with matched non-recipients in Denmark and Finland, 1 October 2024–21 March 2025. Matched seasonal influenza vaccine recipient and non-recipient pairs were followed for a total of 18 weeks after immunisation (defined as 14 days after index date).
The overall VE was 38.1% (31.3–44.8) (Fig. 2). Efluelda Tetra (SV-HD) vaccine had the highest combined VE at 49.9% (26.4–73.4). In Denmark, this corresponded to a RD of −156.6 (−260.4 to −52.8) per 100,000 vaccinated individuals, and in Finland to a RD of 32.8 (−164.0 to 229.6) per 100,000 vaccinated individuals.
Fig. 2.
Risk of laboratory-confirmed influenza A comparing seasonal recipients with non-recipients at day 126 (week 18) in Denmark and Finland, 1 October 2024–21 March 2025.
The VE of Fluad Tetra (SU-SD-Adj) was 40.5% (31.7–49.2). In Denmark, this corresponded to a RD of −233.3 (−274.1 to −192.6) per 100,000 vaccinated individuals, and in Finland to a RD of −244.3 (−352.1 to −136.5) per 100,000 vaccinated individuals. The VE of Vaxigrip Tetra (SV-SD) was 36.9% (30.1 to 43.8). In Denmark, this corresponded to a RD of −155.1 (−207.5 to −102.6) per 100,000 vaccinated individuals, and in Finland to a RD of −112.7 (−137.7 to −87.7) per 100,000 vaccinated individuals.
Lastly, the VE of Influvac Tetra (SU-SD) was 33.0% (−10.8 to 76.8). In Denmark, this corresponded to a RD of −62.4 (−176.1 to 51.2) per 100,000 vaccinated individuals, and in Finland to a RD of −226.3 (−585.8 to 133.2) per 100,000 vaccinated individuals.
Laboratory-confirmed influenza B
Cumulative incidences of laboratory-confirmed influenza B remained very low throughout the season in Denmark and Finland, reflecting its limited circulation during the study period (Fig. 1).
The VE for any influenza vaccine was 63.7% (44.2–83.1) (Fig. 3). The number of events among recipients was limited for this outcome. Vaxigrip Tetra (SV-SD) had a VE of 77.4% (59.0–95.8). The RD was −7.4 (−15.7 to 0.9) per 100,000 vaccinated individuals in Denmark, and −4.9 (−7.9 to −1.8) per 100,000 vaccinated individuals in Finland.
Fig. 3.
Risk of laboratory-confirmed influenza B comparing seasonal recipients with non-recipients at day 126 (week 18) in Denmark and Finland, 1 October 2024–21 March 2025.
Influenza hospitalisation
Influenza-related hospitalisation incidence was lower in Finland than in Denmark. In Denmark, incidence increased steadily, while in Finland it plateaued earlier. Despite these differences, lower incidence among vaccine recipients was seen in both countries (Fig. 1).
The overall estimated VE was 46.8% (40.8–52.9) (Fig. 4). The VE of Efluelda Tetra (SV-HD) was 63.4% (38.1–88.7) in Denmark. This corresponded to a RD of −117.2 (−194.6 to −39.9) per 100,000 vaccinated individuals. The VE and RD in Finland could not be estimated.
Fig. 4.
Risk of influenza hospitalisation comparing seasonal recipients with non-recipients at day 126 (week 18) in Denmark and Finland, 1 October 2024–21 March 2025.
The VE of Fluad Tetra (SU-SD-Adj) was 48.2% (40.8–55.6). In Denmark, this corresponded to a RD of −131.2 (−159.7 to −102.7) per 100,000 vaccinated individuals, and in Finland to a RD of −23.1 (−51.6 to 5.3) per 100,000 vaccinated individuals.
The VE of Vaxigrip Tetra (SV-SD) was 43.6% (23.7–63.6). In Denmark, this corresponded to a RD of −34.3 (−64.1 to −4.5) per 100,000 vaccinated individuals, and in Finland to a RD of −21.5 (−29.3 to −13.6) per 100,000 vaccinated individuals.
The VE of Influvac Tetra (SU-SD) could only be estimated in Denmark and was 30.6% (−7.8 to 69.1) with a RD of −41.9 (−104.5 to 20.8) per 100,000 vaccinated individuals.
Influenza-related death
Vaccine recipients had lower cumulative incidence of influenza-related death when compared to non-recipients, with patterns consistent across Denmark and Finland (Fig. 1).
The overall estimated VE was 63.2% (53.6–72.8) (Fig. 5). The VE of Fluad Tetra (SU-SD-Adj) was 65.8% (54.6–76.9). In Denmark, this corresponded to a RD of −39.7 (−52.4 to −27.0) per 100,000 vaccinated individuals, and in Finland to a RD of −45.2 (−75.4 to −15.1) per 100,000 vaccinated individuals.
Fig. 5.
Risk of influenza-related death comparing seasonal recipients with non-recipients at day 126 (week 18) in Denmark and Finland, 1 October 2024–21 March 2025.
The VE of Vaxigrip Tetra (SV-SD) was 45.2% (−13.9 to 100.0). In Denmark, this corresponded to a RD of 0.5 (−10.6 to 11.6) per 100,000 vaccinated individuals, and in Finland to a RD of −10.0 (−14.9 to −5.1) per 100,000 vaccinated individuals.
Subgroups and waning
Overall, the VE estimates were consistent across age groups and sex, with expected higher RD among individuals aged ≥75 years across the outcomes (Supplementary Table S6). Fig. 6 shows the waning across all brands of seasonal influenza vaccine VE, stratified by 3-week intervals with the per 3-week percentage point change in VE during follow-up estimated by the trend line. The vaccines had an initial VE of 60.1% (43.9–76.3) against laboratory confirmed influenza A, 65.7% (36.9–94.4) against influenza hospitalisation and 74.9% (19.5–100) against influenza-related death at week 3. Subsequently, gradual waning of −7.9 (−10.6 to −5.1), −6.5 (−10.5 to 2.5), and −4.4 (−12.6 to 3.8) percentage points against laboratory confirmed influenza A, hospitalisation, and death, respectively, was observed every 3 weeks (Fig. 6).
Fig. 6.
Waning vaccine effectiveness against laboratory confirmed influenza A, hospitalisation and death related to influenza during 18 weeks of follow-up, comparing recipients of seasonal influenza vaccine with matched non-recipients during 2024/2025 influenza season, stratifying follow-up in 3-week intervals. Waning estimates represent the trend line in the per 3-week vaccine effectiveness estimates.
Supplementary analyses
Test-negative design
We identified 3791 cases (laboratory confirmed influenza A) and 29,763 controls (with confirmed negative tests). Supplementary Table S7 shows the covariate distribution between cases and controls. Among cases and controls, we identified 19,276 recipients of seasonal influenza vaccine and 14,278 non-recipients. The mean age was 78.4 (standard deviation 7.8 years) and 77.8 (standard deviation 8.1) among recipients and non-recipients, respectively (Supplementary Table S8). Among the non-recipients, there were 1672 cases (laboratory confirmed influenza A) and 12,606 controls. Among the vaccine recipients, there were 2119 cases of laboratory-confirmed influenza A, and 17,157 controls. The estimated VE was 38.8% (34.2–43.2), which is similar to our meta-analysed main analysis estimate of 38.1% (31.3–44.8) (Supplementary Table S9). Similarly, the Danish estimate from our main analysis was 41.4% (36.5–46.3).
Negative control outcomes
For diverticulitis and lower back pain, the cumulative incidence was consistently higher among vaccine recipients compared to non-recipients across both countries (Supplementary Figure S4), vaccination was associated with increased risk with negative VEs of −26.7% (−38.0 to −15.5) and −21.4% (−33.7 to −9.1), respectively (Supplementary Table S10). In contrast, clavicle fractures were associated with a small protective effect (VE: 18.2%, 2.3–34.1). The VE against all-cause mortality was 42.2% (34.6–49.9) (Supplementary Table S10). E-values can be found in Supplementary Table S12, indicating that relatively modest confounding could account for the non-null associations in diverticulitis, lower back pain, and clavicle fractures.
Discussion
In this multicohort analysis across two Nordic countries of individuals aged ≥65 years, we observed lower rates of laboratory-confirmed influenza A and B, and influenza-related hospitalisation and death with receipt of a seasonal influenza vaccine compared with no receipt from 1 October 2024 to 21 March 2025. Across all brands, we estimated a VE of 38% against laboratory confirmed influenza A, a VE of 64% against laboratory confirmed influenza B, a VE of 47% against influenza hospitalisation, and a VE of 63% against influenza-related death at 18 weeks follow-up. While our study was not designed as a direct comparison between different influenza vaccine brands, Efluelda Tetra (SV-HD) and Fluad Tetra (SU-SD-Adj) appeared to be associated with higher VEs, but lack of statistical precision and age-based differences in the uptake of vaccine brands precludes strong statements. In Finland, low uptake of Efluelda Tetra and a limited number of events prevented estimation of VE against hospitalisation, and VE estimates against laboratory-confirmed influenza A were imprecise. Our risk differences provide important measures of the influenza burden prevented, and are likely to vary according to the baseline risk among the target group for vaccination. Absolute effects were largest for the more vulnerable subgroups defined by age ≥75 years. We observed that protection afforded against laboratory-confirmed influenza A and influenza-related hospitalisation declined significantly throughout the season, with estimated waning of 7.9 and 6.5 percentage points per 3 weeks, as well as an indication of declining protection against death at 4.4 percentage points per 3 weeks. This translates to an estimated waning of 47.4, 39, and 26.4 percentage points by week 18.
The results of our study support the moderate effectiveness of the seasonal influenza vaccines in reducing severe laboratory-confirmed outcomes among individuals aged 65 and older, across two Nordic countries from October 2024 to March 2025. Our findings align well with the available evidence from the current and previous seasons, which are, however, limited to short follow-up time and largely based on TND in contrast to our TTE with survival analysis.4,6,12,13
Our overall VE estimates are in line with recent European studies. A study by Rose et al. from the same 2024–2025 season reported interim VE estimates from eight European studies covering 17 countries during a period up to 31 January 2025.6 Their findings, based on a TND, indicated VE ranging from 40 to 53% in primary care and 34–52% in hospital settings for all ages. Specifically, for Denmark, the estimated VE from hospital settings in their study against all laboratory confirmed influenza was 55% (47–62) among people aged 65 years and older.6 In our study with a study period until 21 March 2025, we observed lower VE when compared to the study by Rose et al.6 We estimated VE of 38.1% (31.3–44.8) among those aged ≥65 years, against laboratory confirmed influenza A, based on our TTE. This may reflect differences in case definitions or study periods. Rose et al. conditioned on sudden onset of symptoms with fever, myalgia and respiratory symptoms among hospitalised patients, compared to only a positive test in our study.6 At day 84, we observed a VE of 47.7% (33.7–61.6) against laboratory-confirmed influenza A, which is closer to the estimate by Rose et al. Moreover, their VE against laboratory confirmed influenza B in primary care in Denmark was 74% (59–83), while we observed 56.6% (27.1–86.2) against influenza B at week 18 in Denmark. Although the VE estimates against influenza B were higher compared to VE against A viruses, circulation of B viruses was limited this season in Denmark.6
A key strength of our study is the availability of brand-specific estimates with both relative and absolute risk differences. This is particularly important given the current lack of evidence on brand-specific influenza VE from the 2024–2025 season. In this context, previous evidence remains informative for interpreting our findings. A systematic review of nine real world evidence studies (7 February 2020–6 September 2021) comprising approximately 53 million participants highlighted the public health importance of adjuvanted and high-dose vaccines in preventing influenza.14 Even though it assessed trivalent rather than quadrivalent vaccines included in our study, the study showed that an adjuvanted trivalent influenza vaccine was more effective than conventional influenza vaccines and equally effective as high-dose influenza vaccine in reducing influenza-related outcomes in adults aged ≥65 years.14
Our findings are also consistent with those of Stuurman et al., who evaluated brand-specific influenza VE as part of the DRIVE project during the 2021–2022 season.5 Despite limited virus circulation and imprecise estimates in older adults, the study demonstrated the feasibility of brand-level VE monitoring across Europe, with VE of Vaxigrip Tetra (SV-SD) reaching up to 81% in some settings. Although conducted in a very different season, DRIVE concluded that brand-specific VE estimation is achievable but often constrained by sample size—a challenge our large, linked Nordic registries are well positioned to address.5 The Nordic health registries and appropriate methodology provide a strong foundation for timely, brand-specific influenza VE studies that support vaccination strategies and inform public health and regulatory decision-making, as detailed in a study report in the HMA-EMA Catalogue.15
This study has several key strengths. It was conducted in large, population-based cohorts across two Nordic countries, plus one Swedish region, using linked registry data, enabling precise estimates and the inclusion of multiple outcomes—laboratory-confirmed influenza A and B, hospitalisation, and death. The use of a structured target trial emulation design and harmonised definitions across settings enhanced internal validity and comparability of results. We report brand-specific VE estimates, as well as estimates by influenza subtype, and translate VE into burden reduction across two age groups, providing a comprehensive picture of vaccine impact in older adults. In addition, the inclusion of a supplementary TND analysis for Denmark allowed for comparison of results across two established methodological approaches, increasing confidence in our findings. However, these findings should be interpreted in light of the study's limitations.
Robust direct comparisons between vaccine products were not the focus of this study as recipient groups differed widely in age, vaccination timing, and geographic coverage, and sample sizes for some products were limited. Ideally, future studies, with larger datasets and more balanced recommendations, could address the relative vaccine effectiveness. We caution against overinterpreting apparent differences between brands due to differences in the target populations.
The statistical precision of brand-specific estimates was limited for less frequently procured vaccines. Precision also depended on infection incidence and vaccine uptake, which varied during the season. In Denmark and Finland, age-specific differences in administered vaccine type could introduce selection bias. In Denmark, people ≥70 received the adjuvanted vaccine, while those aged 65–69 were offered non-adjuvanted vaccines. A subset of 65+ participants took part in a pragmatic trial receiving either Efluelda Tetra (SV-HD) or a standard-dose vaccine.
In Finland, the adjuvanted vaccine was offered free to adults ≥85 years, while others received non-adjuvanted vaccines. Enhanced vaccines were also available for out-of-pocket purchase.
Cohort studies are susceptible to confounding due to differences related to the study outcomes between vaccinated and unvaccinated individuals. The TND analysis, less prone to confounding by healthcare-seeking behaviour, yielded a VE estimate of 38.8%—nearly identical to the TTE estimate of 38.1% (41.4% for Denmark)—providing reassurance against such confounding in the TTE analysis. However, the findings on diverticulitis and lower back pain suggest residual confounding, potentially underestimating vaccine protection. Conversely, the finding on clavicle fractures may indicate a ‘healthy vaccinee effect’, potentially overestimating protection.
We caution against the use of these negative control outcome VE estimates for direct calibration of influenza VE results.10 Such calibration assumes unmeasured confounding factors associated with negative control outcomes are the same as those for influenza, which is questionable given diverging results. Our negative control outcome and all-cause mortality analyses suggest that vaccine recipients and non-recipients differ in factors not included in this study. The degree to which this reflects confounding in influenza-specific estimates is unclear, though bias is likely towards overestimating VE in older adults.
Due to data availability only from Uppsala in Sweden, most of our population was from Denmark and Finland. We had near-real time data until March 21, 2025, from these countries, but only until January 31, 2025, from Sweden. Nevertheless, the inclusion of Sweden demonstrates multi-country analysis feasibility. It is likely Sweden will have national influenza vaccination data in future, though no formal decision has yet been taken. With only two studies contributing to the meta-analyses estimates, it is possible that the between study variance is imprecisely estimated and biased. Meta-analyses results should thus be interpreted cautiously, and in combination with each countries’ individual results when possible.
The timing of vaccination relative to influenza virus circulation could influence VE estimates. Individuals vaccinated earlier or later might have experienced different virus exposure levels, leading to time-related heterogeneity. Our VE estimates should be interpreted within the specific context of the 2024/2025 season.
Conclusions
This multi-country registry-based study provides estimates of influenza VE in individuals aged ≥65 years during the 2024/2025 season in the Nordic region. Seasonal influenza vaccination was moderately effective in reducing the risk of severe laboratory-confirmed outcomes, with Efluelda Tetra (SV-HD) and Fluad Tetra (SU-SD-Adj) appearing to offer favourable protection in their respective target groups. These findings support that enhanced vaccines should be preferred for older adults, if cost-effective. Continued annual monitoring using Nordic health registries remains crucial for informing evidence-based vaccination strategies and regulatory decision-making.
Contributors
Kristyna Faksova contributed to conceptualisation, investigation, methodology, project administration, validation, writing of the original draft, and writing—review and editing. Emilia Myrup Thiesson contributed to conceptualisation, data curation, formal analysis, investigation, methodology, validation, visualisation, and writing—review and editing. Nicklas Pihlström contributed to data curation, formal analysis, investigation, methodology, and validation. Ulrike Baum contributed to data curation, formal analysis, investigation, methodology, supervision, validation, and writing—review and editing. Tor Biering-Sørensen contributed to validation and writing—review and editing. Eero Poukka contributed to data curation, formal analysis, investigation, methodology, validation, and writing—review and editing. Tuija Leino contributed to investigation, validation, and writing—review and editing. Rickard Ljung contributed to investigation, validation, supervision, and writing—review and editing. Anders Hviid contributed to conceptualisation, resources, supervision, validation, and writing—review and editing.
Data sharing statement
Owing to data privacy regulations in each country, the raw data cannot be shared.
Declaration of interests
Anders Hviid, Emilia Myrup Thiesson, and Ulrike Baum declare support from the European Medicines Agency. KF is a fellow of the ECDC Fellowship Programme, supported financially by the European Centre for Disease Prevention and Control (ECDC). The views and opinions expressed herein do not state or reflect those of ECDC. ECDC is not responsible for the data and information collation and analysis and cannot be held liable for conclusions or opinions drawn. EP reports receiving a grant from Finnish Medical Foundation; AH reports unrelated grants from Independent Research Fund Denmark, the Novo Nordisk Foundation, and the Lundbeck Foundation; AH is a scientific board member of VAC4EU; no financial relationships with any organisations that might have an interest in the submitted work in the previous three years; no other relationships or activities that could appear to have influenced the submitted work. TBS has received research grants from Bayer, Novartis, Pfizer, Sanofi Pasteur, GSK, Novo Nordisk, AstraZeneca, Boston Scientific and GE Healthcare, consulting fees from Novo Nordisk, IQVIA, Parexel, Amgen, CSL Seqirus, GSK and Sanofi Pasteur, and lecture fees from AstraZeneca, Bayer, Novartis, Sanofi Pasteur, GE healthcare and GSK.
Acknowledgements
Funding: This research was supported by the European Medicines Agency.
Footnotes
Supplementary data related to this article can be found at https://doi.org/10.1016/j.lanepe.2025.101518.
Appendix A. Supplementary data
References
- 1.Jefferson T., Di Pietrantonj C., Al-Ansary L.A., Ferroni E., Thorning S., Thomas R.E. Vaccines for preventing influenza in the elderly. Cochrane Database Syst Rev. 2010;(2) doi: 10.1002/14651858.CD004876.pub3. https://www.cochranelibrary.com/cdsr/doi/10.1002/14651858.CD004876.pub3/full [cited 2024 Aug 22] Available from: [DOI] [PubMed] [Google Scholar]
- 2.Influenza A og B samlet | luftvejsinfektioner. https://experience.arcgis.com/experience/220fef27d07d438889d651cc2e00076c/page/Influenza-A%2BB/ [cited 2024 Aug 27]. Available from:
- 3.Hergens M.P., Baum U., Brytting M., et al. Mid-season real-time estimates of seasonal influenza vaccine effectiveness in persons 65 years and older in register-based surveillance, Stockholm county, Sweden, and Finland, 2017. Euro Surveill. 2017;22(8) doi: 10.2807/1560-7917.ES.2017.22.8.30469. https://www.eurosurveillance.org/content/10.2807/1560-7917.ES.2017.22.8.30469 [cited 2024 Dec 16] Available from: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Halme J., Syrjänen R.K., Baum U., Palmu A.A. Effectiveness of trivalent influenza vaccines against hospitalizations due to laboratory-confirmed influenza a in the elderly: comparison of test-negative design with register-based designs. Vaccine. 2022;40(31):4242–4252. doi: 10.1016/j.vaccine.2022.05.080. [DOI] [PubMed] [Google Scholar]
- 5.Stuurman A.L., Carmona A., Biccler J., et al. Brand-specific estimates of influenza vaccine effectiveness for the 2021–2022 season in Europe: results from the DRIVE multi-stakeholder study platform. Front Public Health. 2023;11 doi: 10.3389/fpubh.2023.1195409. https://www.drive-eu.org/ [cited 2025 Apr 8] Available from: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Rose A.M.C., Lucaccioni H., Marsh K., et al. Interim 2024/25 influenza vaccine effectiveness: eight European studies, September 2024 to January 2025. Euro Surveill. 2025;30(7):9–19. doi: 10.2807/1560-7917.ES.2025.30.7.2500102. https://www.eurosurveillance.org/content/10.2807/1560-7917.ES.2025.30.7.2500102 [cited 2025 Mar 25] Available from: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Hernán M.A., Wang W., Leaf D.E. Target trial emulation: a framework for causal inference from observational data. JAMA. 2022;328(24):2446. doi: 10.1001/jama.2022.21383. https://jamanetwork.com/journals/jama/fullarticle/2799678 [cited 2024 Jul 5] Available from: [DOI] [PubMed] [Google Scholar]
- 8.Sera F., Armstrong B., Blangiardo M., Gasparrini A. An extended mixed-effects framework for meta-analysis. Stat Med. 2019;38(29):5429–5444. doi: 10.1002/sim.8362. https://pubmed.ncbi.nlm.nih.gov/31647135/ [cited 2025 May 6] Available from: [DOI] [PubMed] [Google Scholar]
- 9.Horne E.M.F., Hulme W.J., Keogh R.H., et al. Waning effectiveness of BNT162b2 and ChAdOx1 covid-19 vaccines over six months since second dose: OpenSAFELY cohort study using linked electronic health records. BMJ. 2022;378 doi: 10.1136/bmj-2022-071249. https://www.bmj.com/content/378/bmj-2022-071249 [cited 2024 Jan 24] Available from: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Lipsitch M., Tchetgen Tchetgen E., Cohen T. Negative controls: a tool for detecting confounding and bias in observational studies. Epidemiology. 2010;21(3):383–388. doi: 10.1097/EDE.0b013e3181d61eeb. https://pubmed.ncbi.nlm.nih.gov/20335814/ [cited 2024 Aug 29] Available from: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.von Elm E., Altman D.G., Egger M., Pocock S.J., Gøtzsche P.C., Vandenbroucke J.P. The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet. 2007;370(9596):1453. doi: 10.1016/S0140-6736(07)61602-X. https://pubmed.ncbi.nlm.nih.gov/18064739/ [cited 2025 Aug 12] Available from: [DOI] [PubMed] [Google Scholar]
- 12.Maurel M., Pozo F., Pérez-Gimeno G., et al. Influenza vaccine effectiveness in Europe: results from the 2022–2023 VEBIS (vaccine effectiveness, burden and impact studies) primary care multicentre study. Influenza Other Respi Viruses. 2024;18(1) doi: 10.1111/irv.13243. https://pmc.ncbi.nlm.nih.gov/articles/PMC10777262/ [cited 2025 Apr 8] Available from: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Maurel M., Howard J., Kissling E., et al. Interim 2023/24 influenza A vaccine effectiveness: VEBIS European primary care and hospital multicentre studies, September 2023 to January 2024. Euro Surveill. 2024;29(8) doi: 10.2807/1560-7917.ES.2024.29.8.2400089. https://www.eurosurveillance.org/content/10.2807/1560-7917.ES.2024.29.8.2400089 [cited 2025 Apr 8] Available from: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Gärtner B.C., Weinke T., Wahle K., et al. Importance and value of adjuvanted influenza vaccine in the care of older adults from a European perspective – a systematic review of recently published literature on real-world data. Vaccine. 2022;40(22):2999–3008. doi: 10.1016/j.vaccine.2022.04.019. [DOI] [PubMed] [Google Scholar]
- 15.Brand-specific influenza vaccine effectiveness in the Nordic countries | HMA-EMA catalogues of real-world data sources and studies. https://catalogues.ema.europa.eu/node/4382/administrative-details [cited 2025 Jun 16]. Available from:
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.