Interim Effectiveness of 2025–2026 mRNA-1283 and BNT162b2 COVID-19 Vaccines Against COVID-19-Related Outcomes Among Adults Aged ≥ 65 Years in the United States Authors: Nevena Vicic, Alina Bogdanov, Heather Hensler, Taylor Ryan, Ni Zeng, Ekkehard Beck, Emily Patry, Machaon Bonafede, Andre B. Araujo, Amanda Wilson Published: 2026 ABSTRACT IntroductionThe 2025–2026 COVID-19 vaccine season introduced updated formulations targeting the LP.8.1 lineage. This interim analysis assessed absolute vaccine effectiveness (aVE) of mRNA-1283 and BNT162b2 on COVID-19 outcomes in adults aged ≥ 65 years.MethodsThis retrospective study used linked electronic health record and administrative claims data through January 31, 2026. Adults ≥ 65 years who received the mRNA-1283 or BNT162b2 2025–2026 COVID-19 vaccine were matched to unvaccinated individuals. Inverse probability of treatment weighting was applied to each vaccine’s matched cohorts to balance covariates. Each vaccine was evaluated independently against its own unvaccinated comparator group. aVE against COVID-19-related hospitalization and medically attended COVID-19 was estimated using Cox proportional hazards models; aVE = 100 × (1 − hazard ratio).ResultsWe identified 233,072 mRNA-1283 recipients and 422,610 BNT162b2 recipients ≥ 65 years. aVE (95% confidence interval [CI]) of mRNA-1283 against COVID-19-related hospitalization and medically attended COVID-19 was 59.3% (39.0%, 72.9%) and 42.0% (35.0%, 48.3%) among adults ≥ 65 years and 66.9% (45.9%, 79.8%) and 50.2% (42.1%, 57.2%) in ≥ 75 years, respectively. The aVE (95% CI) of BNT162b2 against COVID-19-related hospitalization and medically attended COVID-19 was 48.3% (32.4%, 60.5%) and 41.2% (36.2%, 45.8%) in ≥ 65 years and 45.9% (26.0%, 60.4%) and 44.0% (37.8%, 49.6%) in ≥ 75 years, respectively.ConclusionThis study provides the first real-world evidence that mRNA-1283 was associated with protection against COVID-19-related hospitalization and medically attended COVID-19 among vulnerable older adults at highest risk of severe disease. These findings support mRNA-1283 as an important public health tool for reducing the continual burden of COVID-19 in this population.Supplementary InformationThe online version contains supplementary material available at 10.1007/s40121-026-01395-4. Key Summary Points Why carry out this study?Adults aged ≥ 65 years remain at highest risk of severe COVID-19 outcomes despite the availability of updated seasonal vaccines.Real-world effectiveness data for 2025–2026 LP.8.1-targeting mRNA-1283 were not yet available in older adults in the United States.What was learned from this study?In adults aged ≥65 years, both mRNA-1283 and BNT162b2 were associated with significant protection against COVID-19–related hospitalization and medically attended COVID-19.Adjusted aVE against COVID-19–related hospitalization was 59.3% for mRNA-1283 and 48.3% for BNT162b2, while aVE against medically attended COVID-19 was 42.0% and 41.2%, respectively. Protection was also observed in adults ≥75 years.These data support continued uptake of updated COVID-19 vaccines in vulnerable older adults. Why carry out this study?Adults aged ≥ 65 years remain at highest risk of severe COVID-19 outcomes despite the availability of updated seasonal vaccines.Real-world effectiveness data for 2025–2026 LP.8.1-targeting mRNA-1283 were not yet available in older adults in the United States.What was learned from this study?In adults aged ≥65 years, both mRNA-1283 and BNT162b2 were associated with significant protection against COVID-19–related hospitalization and medically attended COVID-19.Adjusted aVE against COVID-19–related hospitalization was 59.3% for mRNA-1283 and 48.3% for BNT162b2, while aVE against medically attended COVID-19 was 42.0% and 41.2%, respectively. Protection was also observed in adults ≥75 years.These data support continued uptake of updated COVID-19 vaccines in vulnerable older adults. Introduction COVID-19 continues to pose a substantial public health burden, with older adults experiencing the highest rates of severe disease, hospitalization, and death [1]. In the United States, adults aged ≥ 65 years account for the majority of COVID-19-related hospitalizations, reflecting immunosenescence and the high prevalence of underlying medical conditions associated with severe outcomes [1, 2]. Despite the widespread availability of vaccines, this population remains vulnerable due to waning immunity, ongoing viral evolution, and suboptimal vaccine uptake in recent seasons [3, 4]. The continued evolution of SARS-CoV-2 has necessitated annual updates to COVID-19 vaccine formulations to maintain alignment with circulating variants [5]. Real-world evidence from prior seasons has demonstrated that updated COVID-19 vaccines reduce the risk of COVID-19-related hospitalization and severe outcomes [6–10]. Previous estimates of absolute vaccine effectiveness (aVE) of mRNA-1273 formulations (encoding the full length spike protein) in patients aged ≥ 65 years from prior years range from 46%–61% for COVID-19 related hospitalizations and were 38% for medically-attended COVID-19 [6, 9, 10]. For the 2025–2026 season, updated formulations targeting the LP.8.1 lineage of the JN.1 subvariant family were authorized in the United States [11]. Unlike prior COVID-19 vaccines, which utilized the full-length spike protein, mRNA-1283 (mNEXSPIKE) encodes only the receptor binding and N-terminal domains of the spike protein [12], which are key targets of neutralizing antibodies and T-cell responses. The vaccine was approved by the US Food and Drug Administration in May 2025 for the prevention of COVID-19 in adults aged ≥ 65 years and in individuals aged 12–64 years with ≥ 1 underlying condition that puts them at increased risk for severe COVID-19. In contrast, BNT162b2 (COMIRNATY) encodes the full-length spike protein of the LP.8.1 lineage and was also authorized for the 2025–2026 season [13]. No published real-world aVE data were available for mRNA-1283 in the 2025–2026 season. Given that mRNA-1283 and BNT162b2 represented the majority of COVID-19 vaccines administered to adults aged ≥ 65 years in the United States during this season, generating timely real-world evidence is critical to inform clinicians, public health decision-makers, and vaccination programs. To address this gap, this interim analysis sought to characterize the real-world aVE of the updated 2025–2026 mRNA-1283 and BNT162b2 COVID-19 vaccines against COVID-19-related hospitalization and medically attended COVID-19 among adults aged ≥ 65 years in the United States using a harmonized protocol. Note that aVE refers to vaccine effectiveness versus an unvaccinated comparator and no comparative vaccine effectiveness between products were assessed in the current study. Methods Data Source This study used electronic health record (EHR) data from the Veradigm Network EHR linked with administrative healthcare claims from Komodo Health for the period August 23, 2024, through January 31, 2026. The EHR component captures information from ambulatory and outpatient primary care and specialty care settings, while the claims component includes inpatient, outpatient, and pharmacy records. This linked dataset has been described previously for vaccine-related research and has been used in prior studies of COVID-19 epidemiology and vaccine effectiveness [6, 10, 14, 15]. The study was conducted in accordance with the Declaration of Helsinki and applicable guidance for observational research. Because the analysis used de-identified data and involved no direct patient contact, separate ethics committee approval and informed consent were not required. Data access was governed by agreements with the data provider, which included provisions to maintain patient confidentiality and comply with applicable privacy requirements. Participants and Study Design This was a retrospective matched cohort study using a common protocol, data source, eligibility criteria, analytic framework, and endpoints to estimate product-specific aVE among adults aged ≥ 65 years in the United States who were vaccinated with either the mRNA-1283 or the BNT162b2 2025–2026 formulation (vaccinated cohorts) between August 28, 2025, and January 24, 2026 (intake period), or had no evidence of vaccination with any 2025–2026 seasonal COVID-19 vaccine (eligible for the unvaccinated cohorts). Patients were followed through January 31, 2026. For each vaccinated cohort (mRNA-1283 and BNT162b2), vaccinated patients were matched 1:1 to unvaccinated individuals based on age, sex, race, ethnicity, geographic region, week of last healthcare activity before start of intake period, and evidence of vaccination with a prior season’s vaccine during the pre-index period. The index date was defined as the date of vaccination among the vaccinated; among the unvaccinated, the matched vaccinated patient’s index date was assigned as the unvaccinated control patient’s index date. To allow for establishment of vaccine-induced immunity, the cohort entry date (CED) was defined as 7 days following the index date. The pre-index period started on August 23, 2024, the date that the prior season’s COVID-19 vaccines became available + 1 day. Individuals were eligible for study inclusion if they were ≥ 65 years of age at the beginning of the intake period, had evidence of database activity during the intake period, and maintained continuous medical and pharmacy enrollment, allowing gaps of ≤ 45 days, from the start of the pre-index period through CED, irrespective of prior vaccination history. Exclusion criteria included missing or inconsistent information on age, sex, or state of residence; evidence of a COVID-19 diagnosis or COVID-19 treatment from 90 days before the index date through CED; receipt of any COVID-19 vaccine other than the index vaccine within 60 days before the index date or between the index date and CED, defined as 7 days after index; or less than 1 day of follow-up. Unvaccinated individuals were required to have no evidence of COVID-19 vaccination through the later of CED or the end of the intake period. COVID-19 vaccination codes are provided in Supplementary Material Tables S1 and S2. The mRNA-1283 and BNT162b2 analyses were conducted independently, each with its own matched unvaccinated comparator cohort. Due to patient selection criteria, individuals included in one vaccinated cohort could not be included in the other vaccinated cohort or either of the unvaccinated cohorts. Because the mRNA-1283 or the BNT162b2 analyses were conducted independently, an unvaccinated individual could be included in both analyses; however, their index date could differ, as it would have been assigned during matching. Subgroup analyses were conducted among adults aged ≥ 75 years on the index date. Outcomes The primary outcome, COVID-19-related hospitalization, was defined as an inpatient hospitalization claim containing a diagnosis code indicating COVID-19 in any diagnosis position. Because the COVID-19 diagnosis was not required to be the primary diagnosis, these events should be interpreted as hospitalizations with documented COVID-19 rather than hospitalizations adjudicated as being caused by COVID-19. The secondary outcome, medically attended COVID-19, was defined as ≥ 1 claim or record with a diagnosis code indicating COVID-19 in any position or care setting, including inpatient admissions captured in the primary outcome, as well as emergency department visits, urgent care visits, office visits, telemedicine visits, and laboratory records. COVID-19 diagnosis and treatment codes are provided in Supplementary Material Table S3. Follow-up for outcome assessment began on the cohort entry date. Patients were followed until the first occurrence of the outcome or until censoring due to health plan disenrollment, receipt of any COVID-19 vaccine during follow-up, or the end of the study period (Fig. 1).Fig. 1Study design Study design Covariates Baseline demographic characteristics, age, sex, race, ethnicity, insurance type, and US geographic region, were captured on the index date. Centers for Disease Control and Prevention defined clinical characteristics and comorbidities associated with increased risk of severe COVID-19 outcomes were assessed during the 12-month pre-index period. These included asthma, cancer, cerebrovascular disease, chronic kidney disease, chronic lung disease, chronic liver disease, cystic fibrosis, diabetes mellitus, including type 1 and type 2 diabetes, disabilities, heart conditions, HIV, mental health disorders, neurologic conditions, obesity, primary immunodeficiencies, pregnancy, physical inactivity, tobacco use, solid organ or hematopoietic cell transplantation, tuberculosis, and use of immunosuppressive medications [2]. Other baseline variables included days since most recent COVID-19 vaccination, days since most recent SARS-CoV-2 infection, weeks since most recent healthcare activity before the start of the intake period, receipt of a prior-season 2024–2025 COVID-19 vaccine, influenza vaccination during the pre-index period, influenza diagnosis during the pre-index period, and healthcare utilization during the pre-index period. The number of individuals who received an influenza vaccine on the same day as the COVID-19 vaccine was summarized descriptively. High-risk and immunocompromising conditions were evaluated from up to 365 days before the index date through the index date, inclusive. Pregnancy was assessed from the index date through 301 days after the index date, inclusive. Healthcare resource utilization was measured from up to 365 days before the index date through 1 day before the index date. Influenza diagnosis and influenza vaccination history were assessed from the start of the pre-index period through 1 day before the index date. COVID-19 vaccination history was evaluated from the start of the pre-index period through 60 days before the index date, and prior COVID-19 history was assessed from the start of the pre-index period through 90 days before the index date. Supplementary Material Table S4 captured with complete list of covariates. Statistical Analysis The mRNA-1283 and BNT162b2 analyses were performed separately. This included the matching procedures and statistical analyses for the overall and age-stratified cohorts. To address potential residual differences in measured baseline confounders after matching, logistic regression models were used to estimate propensity scores and derive inverse probability of treatment weights (IPTWs). Stabilized IPTWs were calculated as the marginal probability of receiving the index vaccine divided by the propensity score for vaccinated individuals, and as the marginal probability of remaining unvaccinated divided by 1 minus the propensity score for unvaccinated individuals. Covariate balance before and after IPTW was evaluated using standardized mean differences (SMDs), with absolute SMD values > 0.1 considered indicative of imbalance. Covariates are summarized descriptively before and after weighting using means and standard deviations for continuous variables and counts and percentages for categorical variables. Cumulative incidence and unadjusted hazard ratios (HRs) were estimated in the unweighted sample, while adjusted HRs were estimated in the weighted sample. Unadjusted HRs were derived from Cox regression models with receipt of the index vaccine as the only predictor. Adjusted HRs were estimated using weighted Cox regression models that included receipt of the index vaccine. Additional adjustment for covariates with residual imbalance after weighting, defined as an SMD > 0.1, was planned; however, all covariates were balanced following IPTW. Cumulative incidence plots with 95% confidence intervals were generated to visually evaluate the proportional hazards assumption. Unadjusted and adjusted vaccine effectiveness for each outcome were then calculated as (1 − HR) × 100% and reported with corresponding 95% confidence intervals. Data cleaning and analytic file generation was conducted via SQL. Statistical analyses were conducted using SAS Version 9.4. Missing data was not imputed. Results Study Population and Baseline Characteristics The final study population included 233,072 mRNA-1283 recipients and 422,610 BNT162b2 recipients aged ≥ 65 years who met the patient selection criteria pre-weighting (Supplementary Material Fig. S1). Baseline characteristics were similar across the mRNA-1283 and BNT162b2 vaccinated cohorts (Supplementary Material Tables S5 and S6). In the mRNA-1283 analysis, the mean (standard deviation [SD]) age of vaccinated patients was 76 (6.4) years, and 56.2% were female. In the BNT162b2 analysis, the mean (SD) age of vaccinated patients was 76 (6.4) years, and 56.3% were female. Most patients in both cohorts had documented COVID-19 vaccination during the prior season (mRNA-1283: 84.7%; BNT162b2: 83.2%). More than 90% had their most recent documented activity with the healthcare system within 4 weeks prior to the index date (mRNA-1283: 91.9%; BNT162b2: 92.2%). As these characteristics were all matching criteria, rates were similar in the unvaccinated cohorts that were matched to each vaccinated cohort. Most vaccinations during the study period occurred during October (42%) or September (38%), consistent across both cohorts. The most common underlying medical conditions pre-weighting in the vaccinated cohort were heart conditions (mRNA-1283: 31.7%; BNT162b2: 33.1%), diabetes mellitus (30.7%; 32.9%, respectively), and obesity (26.6%; 28.3%, respectively); corresponding rates in the unvaccinated cohorts were 35.3%, 35.8%, and 29.3% for the mRNA-1283 analysis and 36.0%, 36.5%, and 29.7% for the BNT162b2 analysis. After weighting, baseline characteristics were well balanced between vaccinated and unvaccinated cohorts in both vaccine analyses, with all SMDs ≤ 0.10. The median (interquartile range) duration of follow-up was 69 (42–95) days for patients vaccinated with mRNA-1283 and 70 (44–94) days for patients vaccinated with BNT162b2 (Supplementary Material Table S7). In the mRNA-1283 cohort, 58.6% aged ≥ 65 years and 56.6% aged ≥ 75 years were vaccinated against influenza on the same day that they received the COVID-19 vaccine; in the BNT162b vaccinated cohort, the proportions were 60.3% and 58.2%, respectively. Absolute Vaccine Effectiveness Against COVID-19-Related Hospitalization and Medically Attended COVID-19 In the mRNA-1283 analysis, a total of 32 COVID-19-related hospitalizations and 466 cases of medically attended COVID-19 were identified pre-weighting in those who were vaccinated, and 83 and 779 in those who were unvaccinated, respectively (Tables 1 and 2). In the BNT162b2 analysis, a total of 74 COVID-19-related hospitalizations and 931 cases of medically attended COVID-19 were identified pre-weighting in those who were vaccinated, and 158 and 1,540 in those who were unvaccinated, respectively. Cumulative incidence curves for hospitalizations and medically attended COVID-19 are presented in Fig. 2.Table 1Incidence and vaccine effectiveness against COVID-19-related hospitalization and medically attended COVID-19 for the mRNA-1283 analysisVaccinated incidence, n (%)Unvaccinated incidence, n (%)Unadjusted HR (95% CI)Adjusted HR (95% CI)Adjusted aVE (95% CI) ≥ 65 yearsCOVID-19-related hospitalization32 (0.01)83 (0.04)0.368 (0.245, 0.554)0.407 (0.271, 0.610)59.3% (39.0%, 72.9%)Medically attended COVID-19466 (0.20)779 (0.33)0.573 (0.511, 0.642)0.580 (0.517, 0.650)42.0% (35.0%, 48.3%) ≥ 75 yearsCOVID-19-related hospitalization21 (0.02)64 (0.05)0.312 (0.191, 0.511)0.331 (0.202, 0.541)66.9% (45.9%, 79.8%)Medically attended COVID-19254 (0.20)497 (0.39)0.487 (0.419, 0.567)0.498 (0.428, 0.579)50.2% (42.1%, 57.2%)Incidence values and unadjusted HRs are pre-weighting; adjusted HRs and aVE estimates are derived from inverse probability of treatment weighted Cox modelsaVE absolute vaccine effectiveness; CI confidence interval; HR hazard ratio; IQR interquartile rangeTable 2Incidence and vaccine effectiveness against COVID-19-related hospitalizations and medically attended COVID-19 for the BNT162b2 analysisVaccinated incidence,n (%)Unvaccinated incidence,n (%)Unadjusted HR (95% CI)Adjusted HR (95% CI)Adjusted aVE (95% CI) ≥ 65 yearsCOVID-19-related hospitalization74 (0.02)158 (0.04)0.441 (0.335, 0.582)0.517 (0.395, 0.676)48.3% (32.4%, 60.5%)Medically attended COVID-19931 (0.22)1540 (0.36)0.572 (0.528, 0.621)0.588 (0.542, 0.638)41.2% (36.2%, 45.8%) ≥ 75 yearsCOVID-19-related hospitalization57 (0.03)111 (0.05)0.483 (0.351, 0.665)0.541 (0.396, 0.740)45.9% (26.0%, 60.4%)Medically attended COVID-19546 (0.24)955 (0.42)0.541 (0.487, 0.602)0.560 (0.504, 0.622)44.0% (37.8%, 49.6%)Incidence values and unadjusted HRs are pre-weighting; adjusted HRs and aVE estimates are derived from inverse probability of treatment weighted Cox modelsaVE absolute vaccine effectiveness; CI confidence interval; HR hazard ratio; IQR interquartile rangeFig. 2Cumulative incidence curves of time to event with patients censored at the earliest of outcome event, lost to follow-up, or end of data (January 31, 2026). a COVID-19-related hospitalizations: mRNA-1283 analysis. b Medically attended COVID-19: mRNA-1283 analysis. c COVID-19-related hospitalizations: BNT162b2 analysis. d Medically attended COVID-19: BNT162b2 analysis Table 1Incidence and vaccine effectiveness against COVID-19-related hospitalization and medically attended COVID-19 for the mRNA-1283 analysisVaccinated incidence, n (%)Unvaccinated incidence, n (%)Unadjusted HR (95% CI)Adjusted HR (95% CI)Adjusted aVE (95% CI) ≥ 65 yearsCOVID-19-related hospitalization32 (0.01)83 (0.04)0.368 (0.245, 0.554)0.407 (0.271, 0.610)59.3% (39.0%, 72.9%)Medically attended COVID-19466 (0.20)779 (0.33)0.573 (0.511, 0.642)0.580 (0.517, 0.650)42.0% (35.0%, 48.3%) ≥ 75 yearsCOVID-19-related hospitalization21 (0.02)64 (0.05)0.312 (0.191, 0.511)0.331 (0.202, 0.541)66.9% (45.9%, 79.8%)Medically attended COVID-19254 (0.20)497 (0.39)0.487 (0.419, 0.567)0.498 (0.428, 0.579)50.2% (42.1%, 57.2%)Incidence values and unadjusted HRs are pre-weighting; adjusted HRs and aVE estimates are derived from inverse probability of treatment weighted Cox modelsaVE absolute vaccine effectiveness; CI confidence interval; HR hazard ratio; IQR interquartile range Incidence and vaccine effectiveness against COVID-19-related hospitalization and medically attended COVID-19 for the mRNA-1283 analysis Incidence values and unadjusted HRs are pre-weighting; adjusted HRs and aVE estimates are derived from inverse probability of treatment weighted Cox models aVE absolute vaccine effectiveness; CI confidence interval; HR hazard ratio; IQR interquartile range Table 2Incidence and vaccine effectiveness against COVID-19-related hospitalizations and medically attended COVID-19 for the BNT162b2 analysisVaccinated incidence,n (%)Unvaccinated incidence,n (%)Unadjusted HR (95% CI)Adjusted HR (95% CI)Adjusted aVE (95% CI) ≥ 65 yearsCOVID-19-related hospitalization74 (0.02)158 (0.04)0.441 (0.335, 0.582)0.517 (0.395, 0.676)48.3% (32.4%, 60.5%)Medically attended COVID-19931 (0.22)1540 (0.36)0.572 (0.528, 0.621)0.588 (0.542, 0.638)41.2% (36.2%, 45.8%) ≥ 75 yearsCOVID-19-related hospitalization57 (0.03)111 (0.05)0.483 (0.351, 0.665)0.541 (0.396, 0.740)45.9% (26.0%, 60.4%)Medically attended COVID-19546 (0.24)955 (0.42)0.541 (0.487, 0.602)0.560 (0.504, 0.622)44.0% (37.8%, 49.6%)Incidence values and unadjusted HRs are pre-weighting; adjusted HRs and aVE estimates are derived from inverse probability of treatment weighted Cox modelsaVE absolute vaccine effectiveness; CI confidence interval; HR hazard ratio; IQR interquartile range Incidence and vaccine effectiveness against COVID-19-related hospitalizations and medically attended COVID-19 for the BNT162b2 analysis Incidence values and unadjusted HRs are pre-weighting; adjusted HRs and aVE estimates are derived from inverse probability of treatment weighted Cox models aVE absolute vaccine effectiveness; CI confidence interval; HR hazard ratio; IQR interquartile range Cumulative incidence curves of time to event with patients censored at the earliest of outcome event, lost to follow-up, or end of data (January 31, 2026). a COVID-19-related hospitalizations: mRNA-1283 analysis. b Medically attended COVID-19: mRNA-1283 analysis. c COVID-19-related hospitalizations: BNT162b2 analysis. d Medically attended COVID-19: BNT162b2 analysis In the primary analysis of adults aged ≥ 65 years, both vaccines demonstrated significant aVE against COVID-19-related hospitalization and medically attended COVID-19 (Tables 1 and 2). After IPTW adjustment, the aVE (95% CI) against COVID-19-related hospitalization was 59.3% (39.0%, 72.9%) with mRNA-1283 and 48.3% (32.4%, 60.5%) with BNT162b2. The adjusted aVE (95% CI) against medically attended COVID-19 was 42.0% (35.0%, 48.3%) with mRNA-1283 and 41.2% (36.2%, 45.8%) with BNT162b2. Age Subgroup Analyses In the mRNA-1283 analysis, the vaccinated cohort was comprised of 126,136 adults aged ≥ 75 years pre-weighting. Baseline characteristics are reported in Supplementary Material Table S8. Among adults aged ≥ 75 years, the aVE against COVID-19-related hospitalization was 66.9% (45.9%, 79.8%) and the aVE against medically attended COVID-19 was 50.2% (42.1%, 57.2%). In the BNT162b2 analysis, the vaccinated cohort was comprised of 227,989 adults aged ≥ 75 years pre-weighting. Baseline characteristics are reported in Supplementary Material Table S9. Among adults aged ≥ 75 years, the aVE against COVID-19-related hospitalization was 45.9% (26.0%, 60.4%) and the aVE against medically attended COVID-19 was 44.0% (37.8%, 49.6%). Discussion In this analysis of a large linked EHR and administrative claims dataset, both the 2025–2026 mRNA-1283 and BNT162b2 COVID-19 vaccines demonstrated significant protection against COVID-19-related hospitalization and medically attended COVID-19 among adults aged ≥ 65 years in the United States. The aVE against COVID-19-related hospitalization in this population was 59.3% for mRNA-1283 and 48.3% for BNT162b2, whereas the aVE against medically attended COVID-19 was 42.0% and 41.2%, respectively.. Because each vaccine cohort was evaluated independently against its own unvaccinated comparator, these estimates should be interpreted as vaccine-specific measures of effectiveness rather than as evidence of comparative effectiveness between products. This study provides the first real-world evidence of mRNA-1283 effectiveness. The targeted design of mRNA-1283 encodes only the receptor binding and N-terminal domains of the spike protein. This allows for lower mRNA doses compared with COVID-19 vaccines that encode the full-length spike (10 µg for mRNA-1283 vs 50 µg for mRNA-1273 and 30 µg for BNT162b2) [13, 16, 17]. In the phase 3 NextCOVE trial, mRNA-1283 elicited robust immune responses and demonstrated higher efficacy point estimates compared to mRNA-1273 in older adults (13.5%, 95% CI, − 7.7 to 30.6), including individuals at increased risk of severe COVID-19 (28.1%, 95% CI, 4.4 to 45.9) [18]. These findings support the relevance of evaluating mRNA-1283 effectiveness in older adult populations in real-world settings. To contextualize the potential public health implications of the observed real-world effectiveness estimates, a published economic evaluation of mRNA-1283 in the United States by Fust et al. [19] was adapted using the aVE estimates from the present study. In an analysis with the COVID-19 disease incidence inputs and other model parameters from the original model, mRNA-1283 was estimated to prevent approximately 117,000 hospitalizations and 15,285 deaths among adults aged ≥ 65 years compared with no COVID-19 vaccination. These modeled estimates provide context for the potential population-level impact of the observed protection against COVID-19 related hospitalizations in an older adult population. Although the present analysis applied a harmonized protocol to estimate aVE for both mRNA-1283 and BNT162b2, it was not designed or powered as a head-to-head comparative effectiveness study, and no formal statistical comparison between vaccine products was performed. Therefore, the numerically higher aVE point estimates observed for mRNA-1283 for COVID-19-related hospitalization, including among adults aged ≥ 75 years, should be interpreted cautiously and considered descriptive only. Further evaluation in additional datasets and populations, with longer follow-up and analytic designs specifically intended for direct product comparisons, is needed to characterize durability and support future comparative effectiveness analyses. Limitations Although we have identified and adjusted for many baseline confounders, cohort assignment was not randomized, and residual confounding may remain, as is inherent to observational study designs. For instance, vaccinated individuals may engage in other behaviors associated with a healthier lifestyle, such as more frequent exercise, contributing to their lower rates of hospitalization with COVID-19 and this may not be addressed with the available covariates. Also, the study population only includes individuals engaged with the healthcare system, possibly excluding others due to barriers such as trust, socioeconomic status, or accessibility. The outcome definition of hospitalization with COVID-19 partially mitigates this limitation, as severe outcomes are less influenced by variation in healthcare-seeking behavior. Although the study design requires that vaccinated individuals remain event-free for 7 days following vaccination, the use of a symmetric study design among the unvaccinated cohorts minimizes the introduction of bias between cohorts. As with all claims and/or EHR data, vaccination status may be incompletely captured, which could have led to bias toward the null hypothesis and resulted in underestimation of true aVE. In addition, outcomes may be underestimated if people avoid obtaining medical care. While treatment avoidance is possible in both vaccinated and unvaccinated individuals, patients who avoid vaccination may be more likely to avoid other medically-necessary care. This difference in healthcare seeking behavior would also bias towards underestimating aVE Misclassification of both exposure and outcomes is also possible, as diagnostic and procedural codes used to define vaccination and COVID-19-related hospitalization may be subject to coding errors. However, validation studies suggest that outcome misclassification is likely limited; for example, Kadri et al. reported that the ICD-10-CM code U07.1 for COVID-19 had a positive predictive value of approximately 98% when compared with clinical documentation and laboratory-confirmed infection [20]. A further limitation is that for some patients COVID-19 may be an incidental diagnosis that is captured as part of a hospitalization for an unrelated health event. While this is possible, the weighting process should ensure that the odds of an unrelated hospitalization are similar between cases and controls. Therefore, while this may lead to an overestimate of the frequency of COVID-19-related hospitalizations it should not bias the aVE results. Furthermore, the low frequency of hospitalizations, particularly in the subgroup analysis, resulted in large confidence intervals in adjusted aVE. Additional analyses with data from the full season that allows for the capture of more events over a longer follow-up time may help assess the durability of vaccine effectiveness and improve confidence in these findings. Additionally, these data sources may not fully capture all hospitalizations, particularly among populations with limited or fragmented access to care, such as uninsured individuals, underrepresented groups, or those receiving care outside the insurance network. The patient selection also excluded patients based on receipt of immunomodulators used to treat COVID-19; however, this approach likely excluded individuals who received these medications for non-COVID-19 indications. This may affect generalizability and introduce selection bias, though the exclusions were applied to both cases and controls. Strengths This retrospective observational cohort study used both matching and IPTW to address potential confounding arising from baseline differences between recently vaccinated and unvaccinated cohorts. Index dates for individuals in the unvaccinated cohorts were assigned through direct matching on demographic and clinical characteristics, including age, sex, race, ethnicity, geographic region, week of most recent healthcare activity before the intake period, and receipt of a 2024–2025 COVID-19 vaccine. Matching on recent healthcare activity was included to reduce the potential for surveillance bias by balancing healthcare-seeking behavior across cohorts. IPTW was then applied to further account for measured confounding and to estimate the average treatment effect in the overall study population. Despite limited sample size, the use of a harmonized protocol, data source, eligibility criteria, analytic framework, and endpoints ensured that data from individuals in each group, vaccinated and unvaccinated, were handled the same. Conclusion In this analysis of data on the effectiveness of the 2025–2026 mRNA COVID-19 vaccines, both mRNA-1283 and BNT162b2 demonstrated substantial aVE against COVID-19-related hospitalization and medically attended COVID-19 among adults aged ≥ 65 years in the United States. Among patients vaccinated with mRNA-1283, numerically higher aVE was observed in patients ≥ 65 years of age against COVID-19-related hospitalization though this trend needs to be verified with additional analyses. These findings provide the first real-world evidence of mRNA-1283 effectiveness in older adults, support continued uptake of updated COVID-19 vaccines, and provide direction for future comparative vaccine effectiveness studies. Supplementary Information Below is the link to the electronic supplementary material.Supplementary file1 (PDF 1270 KB) Supplementary file1 (PDF 1270 KB)