XBB Boosters Keep Neutralizing Antibodies Detectable at 12 Months in Older Adults
Older adults maintained detectable neutralizing antibodies one year after receiving one of three XBB-containing protein boosters, but the quality and durability of that response varied by formulation and by whether participants became infected during follow-up. The cohort included 90 adults aged 65 years and older, extending earlier 6-month observations into a 12-month analysis of persistence, decay kinetics, and protection against SARS-CoV-2 infection.
Across all three vaccine groups, neutralizing activity remained measurable at day 357, and geometric mean titers were broadly comparable at that time point. In the Tri-XBB.1.5 cohort, day-357 titers ranged from 313.1 against XEC to 3,583.0 against BA.5. In the Bi-Omi-XBB cohort, titers ranged from 452.6 to 2,047.0, while the Tetra-XBB.1 cohort ranged from 338.0 against XEC to 1,915.0 against BA.5. Even with those different starting points and variant-specific slopes of decline, the end-of-year titers converged across the three arms.
The pattern fits the immunology of aging in a straightforward way: older adults often mount responses that are detectable but not always equally durable, and a preserved titer does not necessarily mean identical protection across circulating variants. Vaccine composition, antigenic match, and the way memory B cells are recalled appear to shape what remains useful a year later.
Infection Events Briefly Lifted Titers, Then the Advantage Faded
During the 6- to 12-month follow-up period, seven confirmed symptomatic infections and seven probable infections were documented. Those breakthrough events temporarily boosted antibody levels, especially against JN.1 and XEC, but the effect did not persist as a clear long-term advantage. By the end of follow-up, antibody levels had converged again, whether infection occurred early, late, or during the second half of the year.
Two things happened at once. First, infection acted like a natural antigenic recall, briefly increasing titers in some participants. Second, the timing of that infection altered the decay curve itself. Early-infected participants showed significant declines against nearly all variants except KP.2, with the largest reductions against BA.5 and XBB.1.5. Late-infected participants also showed broad declines, although JN.1 and KP.3 were spared from statistically significant reduction in that subgroup. Participants infected during the 6- to 12-month interval showed rising titers against JN.1 and XEC, while D614G and BA.5 declined.
The biologic message is blunt. Breakthrough infection can lift antibody levels, but it does not reliably confer a durable ceiling of protection, especially when the viral target keeps moving. Antigenic drift continues to narrow the distance between prior vaccine targets and current variants, and the study’s cartography data captured that shrinkage in a way serology alone cannot.
Protection Against Infection Favored Tri-XBB.1.5 Over Bi-Omi-XBB
The most clinically consequential result was the difference in infection risk between formulations. Compared with Bi-Omi-XBB, the Tri-XBB.1.5 vaccine reduced confirmed infection risk by 80%, with a hazard ratio of 0.20 and a 95% confidence interval of 0.05 to 0.79. The Tetra-XBB.1 vaccine reduced confirmed infection risk by 72%, with a hazard ratio of 0.28 and a 95% confidence interval of 0.09 to 0.90. When probable infections were added to the endpoint, Tri-XBB.1.5 still showed a significant 73% risk reduction, while the Tetra-XBB.1 estimate no longer reached conventional statistical significance.
The infection-free survival curves separated around day 92 and stayed apart through follow-up. That timing matters clinically because it suggests the strongest divergence emerged after the first few months rather than immediately after boosting, when antibody titers were still comparatively high across all groups. In other words, the protection gap was not simply a short-lived peak response; it persisted as titers waned and variant pressure accumulated.
This is where neutralization data and real-world effectiveness begin to part ways. Comparable day-357 titers did not translate into identical infection protection. The likely explanation is functional immunity beyond concentration alone: antibody avidity, epitope targeting, and Fc-mediated activity may all influence whether a measured titer actually stops infection at the mucosal surface. A vaccine that better matches the circulating strain may also generate antibodies that bind more efficiently to relevant epitopes, even if total titers eventually look similar on paper.
Decay Kinetics Suggest the Vaccine Platform Shapes Durability
The decay analyses showed distinct half-lives across formulations and variants. Tri-XBB.1.5 produced the most durable response against D614G, with an estimated half-life of 622 days, while its decay against XEC was much faster, with a half-life of 143 days. Bi-Omi-XBB showed slower decay against several emerging variants, including half-lives of 466 days for XBB.1.5, 484 days for JN.1, 384 days for KP.2, 527 days for KP.3, and 984 days for XEC. Tetra-XBB.1 sat between those patterns, with half-lives of 362 days for JN.1, 420 days for KP.2, 352 days for KP.3, and 717 days for XEC, while early variants decayed more quickly.
Among early-infected participants, decay against later variants followed a slower trajectory, with especially long half-lives for XEC and JN.1. That suggests infection itself reshapes the maturation of the antibody response, likely by broadening memory B-cell recall and changing the breadth of the antibody pool. Older adults who experienced infection after boosting may therefore have gained a different quality of immune imprint than those who remained uninfected, even when the final titers converged.
The antigenic mapping reinforced that point. Distances from D614G to newer variants narrowed over time, but the magnitude of that contraction differed by vaccine group and infection status. XEC showed the greatest contraction in some groups, while KP.3 and XEC moved closer in others. Those shifts are consistent with a moving viral target and a host response trying to keep pace.
Practical Limits of the Data Leave Several Questions Open
Several limitations narrow how far these findings can be pushed. The cohort was modest in size, especially for infection-incidence analyses, and subgroup hazard ratios carried wide confidence intervals. The study measured neutralizing antibodies, but not cellular immunity, Fc-mediated effector functions, or epitope mapping. Probable infections were inferred from serologic evidence, which introduces some uncertainty even though the confirmed-infection analysis pointed in the same direction.
Virologic screening was not systematic, viral sequencing was not performed, and acute-phase samples were not collected after infection. As a result, the specific infecting variants remain unknown, and it is not possible to cleanly separate antibody waning from infection-driven boosting in every participant. Recently emerged lineages such as BA.3.2 were also outside the variant-testing panel.
Even with those gaps, the clinical signal is hard to ignore: all three XBB protein boosters maintained measurable neutralization for 12 months in older adults, but Tri-XBB.1.5 provided stronger protection against confirmed infection than Bi-Omi-XBB in this cohort. The remaining question is not whether antibodies persist—they do—but which vaccine composition best preserves meaningful protection as SARS-CoV-2 keeps evolving.