Comparative effectiveness of heterologous booster schedules with AZD1222, BNT162b2, or mRNA-1273 vaccines against COVID-19 during omicron predominance in the Nordic countries

Abstract

Objective: To investigate the effectiveness of heterologous booster schedules with AZD1222 (Oxford-AstraZeneca, referred to as AZD), BNT162b2 (Pfizer-BioNTech, BNT), and mRNA-1273 (Moderna, MOD) vaccines compared with primary schedules and with homologous mRNA-vaccine booster schedules during a period of omicron predominance. Design: Population-based cohort analyses. Setting: Denmark, Finland, Norway, and Sweden, 27 December 2020 to 28 February 2022. Participants: Adults that had received at least a primary vaccination schedule (ie, two doses) of the AZD, BNT, and/or MOD vaccines during the study period. Main outcome measures: Using the Kaplan-Meier estimator, we compared country-specific risks of SARS-CoV-2 infection and severe COVID-19 outcomes in heterologous booster vaccinated with primary schedule vaccinated (matched analyses) and homologous booster vaccinated (weighted analyses) since emergence of omicron. Results: Heterologous booster schedules improved protection against all outcomes compared with primary schedules, with the largest and most robust effects observed for severe COVID-19. Risk differences for documented infection ranged from -22.4% to -3.1% (comparative vaccine effectiveness [CVE] 9.7% to 60.9%; >63.2% for COVID-19 hospitalisation) across countries for AZD1BNT2BNT3 (AZD as primary dose followed by two doses of BNT) vs AZD1BNT2 and -22.2% to -3.2% (CVE 37.4% to 67.8%; >34.6% for hospitalisation) for BNT1BNT2MOD3 vs BNT1BNT2, the two most common heterologous booster schedules. Heterologous- and homologous booster schedules had comparable effectiveness. Risk differences of documented infection ranged from -0.4% to 4.4% (CVE -20.0% to 2.4%) for AZD1BNT2BNT3 vs BNT1BNT2BNT3 and -19.8% to 1.7% (CVE -14.6% to 53.8%) for BNT1BNT2MOD3 vs BNT1BNT2BNT3; for most comparisons, risk differences for severe COVID-19 outcomes were smaller than 1 per 1000 vaccinated. Previous infection followed by a booster dose conferred the greatest protection. Conclusion: Heterologous booster vaccine schedules are associated with an increased protection against omicron-related COVID-19 outcomes that is comparable to that afforded by homologous booster schedules.

Design and analysis heterogeneity in observational studies of COVID-19 booster effectiveness: A review and case study

“…We give details on the propensity score generation and matching process in discussion S5. Even less common in the literature was the use of a propensity score not via matching, but with IPW instead ( 17 ). This involved weighting observations in the study sample by the inverse of their propensity for vaccination/booster (more detail is included in tables S10 and S11).…”

“…However, reported VEs for the first booster versus primary series vaccination in such studies have varied considerably, particularly for infection, where estimates from as high as 93% to as low as negative effectiveness can be found (16)(17)(18). Differing epidemiological and clinical landscapes, including preexisting natural immunity, evolution of COVID treatments, and contemporary variant circulation, may explain some of the differences between study results, but there still remains substantial heterogeneity in VE estimates reported by studies in the same country at around the same time, particularly in the US (19,20).…”

Design and Analysis Heterogeneity in Observational Studies of COVID-19 Booster Effectiveness: A Review and Case Study

“…46 Even less common in the literature was the use of a propensity score not via matching, but with IPW instead. 16 This involved weighting observations in the study sample by the inverse of their propensity for vaccination/booster (more detail is included in Table 1). Compared to propensity score matching with a caliper, IPW provides potential efficiency benefits by not fully excluding any observations.…”

“…However, reported VEs for the first booster versus primary series vaccination in such studies have varied immensely, particularly for infection where estimates from as high as 93% to as low as negative effectiveness can be found. [15][16][17] Differing epidemiological circumstances, including pre-existing natural immunity and contemporary variant circulation, may explain some of the differences between study results, but there still remains substantial heterogeneity in VE estimates reported by studies in the same country at around the same time, particularly in the US. 18,19 Another factor that may explain some of the substantial differences in the results of VE studies is heterogeneity in study design and statistical methods.…”