Evaluating the impact of extended dosing intervals on mRNA COVID-19 vaccine effectiveness in adolescents

Tim K. Tsang, Sheena G. Sullivan, Yu Meng, Francisco Tsz Tsun Lai, Min Fan, Xiaotong Huang, Yun Lin, Liping Peng, Chengyao Zhang, Bingyi Yang, Kylie E. C. Ainslie, Benjamin J. CowlingResearch Square 2024 · DOI

TL;DR

This study investigates the effectiveness of extending the dosing interval for mRNA COVID-19 vaccines in adolescents, finding a modest increase in vaccine effectiveness (VE) but also a higher risk of infection between doses. The findings suggest that while extended intervals offer some benefits, they must be weighed against potential risks and other public health considerations.

What Problem Does This Paper Address?

The COVID-19 pandemic saw rapid vaccine development and deployment. For mRNA vaccines like Pfizer-BioNTech's Comirnaty, the standard primary series involved two doses administered within a few weeks. However, some regions extended this interval (e.g., to 28 days or more) for two main reasons: to reduce the risk of myocarditis, a rare side effect more common in younger males after the second dose, and to increase the number of people receiving at least one dose more quickly (dose-sparing). While the myocarditis risk reduction was a key driver, the impact of these extended intervals on vaccine effectiveness (VE) – how well the vaccine prevents infection – was not fully understood, particularly for adolescents. Previous studies often focused only on the period after the second dose, potentially overlooking the increased risk of infection during the longer gap between doses. This paper aims to provide a comprehensive evaluation of the VE of extended dosing intervals in adolescents, considering both the period after the second dose and the inter-dose period.

Methods

The researchers employed a combination of statistical modeling and real-world data analysis.

  • Calendar-time proportional hazards models: This is a standard epidemiological technique used to estimate the risk of an event (in this case, infection) over time, accounting for when individuals entered the study and when they might have been exposed. They used this to compare the hazard of infection between adolescents who received regular (21-27 days) and extended (≥28 days) dosing intervals.
  • Case-density sampling (matching approach): This method selects controls (uninfected individuals) who were at risk of infection at the same time as the cases (infected individuals), helping to control for time-varying factors like epidemic intensity.
  • Simulation studies: To explore the impact of assumptions about vaccine effectiveness waning over time and to validate their methods, the researchers created synthetic datasets. This allowed them to test how different assumptions about protection duration and the rate of waning affected their estimates of hazard ratios.

A key methodological consideration was how to account for "waning VE" – the decrease in vaccine protection over time. They explored different assumptions about the duration of protection and whether it waned perpetually or had a finite end.

Datasets

The study utilized a COVID-19 database from the Center of Health Protection in Hong Kong. This database contained:

  • Vaccination records: Including date and type of vaccine for each dose, demographic information (date of birth, sex), and underlying conditions.
  • Case details: Including confirmed COVID-19 cases (PCR or RAT), notification date, hospitalization, and mortality outcomes.

The analysis focused on adolescents aged 12-17 years in Hong Kong between January and April 2022, a period marked by the Omicron BA.2 wave. Adolescents with underlying conditions were excluded to avoid confounding factors.

Key Insights and Ideas

  • Extended intervals offer a modest VE boost, but with a trade-off: While extending the dosing interval did show a slight increase in VE after the second dose, this benefit was accompanied by a significantly higher risk of infection during the longer interval between the first and second doses.
  • Waning VE is a critical factor: The way vaccine effectiveness wanes over time heavily influences the estimated benefits of extended dosing intervals. Unrealistic assumptions about perpetual waning can lead to overestimation of protection.
  • Previous studies might have overestimated benefits: The authors suggest that a previously reported higher VE for extended intervals might have been an overestimate due to how waning VE was handled in those analyses.
  • Myocarditis risk reduction is a separate, important consideration: The study acknowledges that the decision to extend intervals is complex and involves balancing VE against the reduced risk of myocarditis.

Key Results

  • VE after second dose: Calendar-time proportional hazards models estimated a hazard ratio (HR) for infection of 0.86 to 0.99 for extended versus regular dosing intervals, translating to an absolute increase in VE of 1-14%. Matching approaches yielded similar results (HR 0.85-0.87).
  • Risk during the inter-dose period: Adolescents in extended dosing groups had a significantly higher hazard of infection during the period between the first and second doses (HR: 1.66; 95% CI: 1.07, 2.59; p = 0.02).
  • Simulation findings: Simulations showed that assuming a finite duration for waning VE was crucial for accurate estimation. Perpetual waning assumptions led to underestimation of the hazard ratio (overestimation of VE).
  • Sensitivity analyses: Varying the definition of an "extended" interval (e.g., 56 days instead of 28) and the assumed VE did not drastically change the core findings, though perpetual waning assumptions consistently led to different results.

Limitations and Future Work

The authors acknowledge several limitations:

  • Observational study design: They cannot completely rule out unidentified confounders, such as differences in health-seeking behavior or eligibility for vaccination that might influence the choice of dosing interval.
  • Projected VE: VE estimates for the primary series were projected backward from waning VE, as no unvaccinated individuals were included for direct comparison.
  • Variant information: Specific variant data for cases was unavailable, though Omicron BA.2 was predominant.
  • Asymptomatic/mild cases: Some milder or asymptomatic infections might have been missed, potentially affecting case counts.
  • Dynamic decision-making: The decision to vaccinate or choose an interval might have changed based on the evolving epidemic.

Future work could involve further exploring the optimal balance between VE, myocarditis risk, and dose-sparing, especially as new variants emerge and booster strategies evolve. The study also highlights the critical need for robust methods to account for VE waning in future vaccine effectiveness studies.

Summary updated

Abstract

Extending the dosing interval of a primary series of mRNA COVID-19 vaccination has been employed to reduce myocarditis risk in adolescents, but previous evaluation of impact on vaccine effectiveness (VE) is limited to risk after second dose. Here, we quantified the impact of the dosing interval based on case notifications and vaccination uptake in Hong Kong from January to April 2022. We estimated that the hazard ratio (HR) and odds ratio (OR) of infections after the second dose for extended (28 days or more) versus regular (21-27 days) dosing intervals ranged from 0.86 to 0.99 from calendar-time proportional hazards models, and from 0.85 to 0.87 from matching approaches, respectively. Adolescents in the extended dosing groups (including those who did not receive a second dose in the study period) had a higher hazard of infection than those with a regular dosing interval during the intra-dose period (HR: 1.66; 95% CI: 1.07, 2.59; p = 0.02) after the first dose. Implementing an extended dosing interval should consider multiple factors including the degree of myocarditis risk, the degree of protection afforded by each dose, and the extra protection achievable using an extended dosing interval.
BibTeX
@inproceedings{tsang2024evaluating,
  author = {Tim K. Tsang and Sheena G. Sullivan and Yu Meng and Francisco Tsz Tsun Lai and Min Fan and Xiaotong Huang and Yun Lin and Liping Peng and Chengyao Zhang and Bingyi Yang and Kylie E. C. Ainslie and Benjamin J. Cowling},
  title = {{Evaluating the impact of extended dosing intervals on mRNA COVID-19 vaccine effectiveness in adolescents}},
  booktitle = {Research Square},
  year = {2024},
  doi = {10.21203/rs.3.rs-4518813/v1},
}
Figure