Vaccine Effectiveness Against Severe Acute Respiratory Syndrome Coronavirus 2 Delta and Omicron Infection and Infectiousness Within Households in the Netherlands Between July 2021 and August 2022

The study investigates two components of COVID-19 vaccine impact: vaccine effectiveness against infection (VE-infection), which reduces susceptibility to becoming infected, and vaccine effectiveness against infectiousness (VE-infectiousness), which reduces onward transmission from vaccinated infected individuals. Household studies are well-suited to assess both endpoints. Understanding both effects is important for vaccination policy: if vaccines mainly reduce susceptibility, targeting vulnerable groups may suffice; if vaccines also reduce infectiousness, broader vaccination (e.g., healthcare workers) can further protect high-risk individuals. Prior research showed substantial protection against infection that varies by variant and wanes over time. The Delta variant exhibited increased transmissibility and became dominant in 2021; Omicron and its subvariants are even more transmissible with immune escape. The objective of this study was to estimate VE-infection and VE-infectiousness for Delta and Omicron in the Netherlands during late 2021 and early-to-mid 2022, using a household substudy within a large prospective cohort.

Existing studies have demonstrated vaccine-induced reductions in infection risk and varying impacts on transmission, with differences across variants and over time since vaccination. Reports indicate higher infectiousness of Omicron compared to Delta, and mixed findings on VE-infectiousness in early Omicron waves (often BA.1). Dutch source- and contact-tracing analyses during Delta found notable reductions in transmission from vaccinated index cases. Other studies (e.g., Jalali et al., Lyngse et al., Tan et al.) provide context on Omicron’s increased transmissibility, immune escape, and differing infectiousness profiles for breakthrough and reinfections. These findings frame the need to quantify both VE-infection and VE-infectiousness in households across variant periods.

Design: VASCO is a prospective cohort of ~45,000 community-dwelling Dutch adults aged 18–85 years (enrolled 3 May–15 Dec 2021; 5-year follow-up). Participants complete monthly questionnaires on COVID-19 vaccination and self-reported positive SARS-CoV-2 tests (PCR or self-administered antigen). Dates and test types are recorded via a study app. From April 2022, self-test kits were provided free of charge due to the cessation of free national testing. Starting late August 2021, participants with a positive test were asked to complete, one month post-test, a questionnaire about household members, including their age, SARS-CoV-2 testing (within 2 weeks before to 4 weeks after the participant’s positive test), test dates/results, and COVID-19 vaccination history. Verbal permission from household members was recorded. Study period: 23 July 2021 to 7 August 2022, covering dominance of Delta and Omicron (BA.1, BA.2, BA.4/BA.5) in the Netherlands. Dominance was defined as >90% of sequenced samples. Delta-dominant: start–18 Dec 2021; Omicron-dominant: 11 Jan 2022–end; BA.1: 9–31 Jan 2022; BA.2: 20 Mar–16 May 2022; BA.4/5: 27 Jun–end. Inclusion/exclusion: Included households where the VASCO participant (index case) was first to test positive per questionnaire. Excluded if the index had prior confirmed infection before index date or anti-nucleocapsid (anti-N) seropositivity before index date; if co-primary cases occurred (index and member positive within 1 day); if any member did not consent to share data; and for VE analyses, household members without a test in the window period were excluded. A secondary case was a household member with a positive test 2–14 days after the index date. Information on prior infections in household members was unavailable. Vaccination status: Determined for index cases (from self-report and, where consented, the national vaccination registry CIMS; combined if partial) and for household members (self-report). Unvaccinated: no reported/registered vaccination. Primary series: two doses of Comirnaty (BNT162b2), Spikevax (mRNA-1273), or Vaxzevria (ChAdOx1-S) ≥14 days before index date, or one dose of Jcovden (Ad26.COV2-S) ≥28 days before index date. First booster: third dose (Comirnaty/Spikevax/Vaxzevria) ≥7 days before index date; a third dose before 18 Nov 2021 (for severe immunodeficiency) was considered additional primary, not booster. For Jcovden, a second dose >90 days after first, given ≥7 days prior, counted as booster. Second booster analogous with one extra dose. Individuals with other/indeterminate statuses were excluded. Serology: Fingerprick samples were tested with Elecsys anti-N assays (Roche Cobas e801). Qualitative cutoff index values were calibrated to BAU/mL using batch-specific linear calibration; anti-N positivity cutoff derived accordingly. Outcomes and analysis: Secondary attack rates (SARs) were the proportion of household members testing positive after the index date, overall and stratified by vaccination status and variant; compared by chi-square tests. VE-infection (Delta and Omicron periods) was estimated via logistic regression with generalized estimating equations (GEE, exchangeable correlation) to account for household clustering. Models adjusted for household member age (0–17, 18–39, 40–59, ≥60), index age (18–39, 40–59, 60–85), calendar week (categorical), index vaccination status, and household size. VE-infectiousness used analogous GEE logistic models adjusting for index age, household member age, household member vaccination status, calendar week, and household size. Sensitivity analyses: (1) stratified by time since vaccination (<90 vs ≥90 days); (2) included index cases with prior infection to assess its impact on VE-infectiousness; (3) excluded household members <18 years to assess age effects; (4) reduced secondary-case window to 7 days. Analyses used R 4.0.2 (geepack). Ethics: Approved by a medical ethics committee (NL76815.056.21); written informed consent obtained.

- Sample: 3399 index-case questionnaires included, covering 4105 household members. Median index age 61 years (IQR 51–65); median household member age 58 years (IQR 34–65). At index date, index vaccination: 2% unvaccinated, 10% primary, 76% first booster, 12% second booster; household members: 11% unvaccinated, 25% primary, 55% first booster, 9% second booster. - Time to secondary infection: median 4.5 days (IQR 3–6) during Delta; 4.0 days (IQR 3–6) during Omicron. - Secondary attack rates (SAR): Overall similar for Omicron (43% [95% CI, 41–44%]) and Delta (41% [95% CI, 36–47%]); by Omicron subvariant: BA.1 40% (95% CI, 33–47%), BA.2 37% (95% CI, 34–40%), BA.4/5 44% (95% CI, 40–48%). SAR with unvaccinated index cases: Delta 55% (95% CI, 38–72%), Omicron 57% (95% CI, 43–71%). SARs generally lower when index cases were vaccinated. - VE-infection (adjusted): • Delta: 47% (95% CI, −27% to 78%) for primary series among household members; insufficient boosted cases for reliable Delta estimates. • Omicron: −36% (95% CI, −87% to 1%) for primary series; −28% (95% CI, −77% to 7%) for first booster. Second booster not reliably estimable due to age imbalance. - VE-infectiousness (adjusted): • Delta: 70% (95% CI, 28% to 87%) for primary series; 93% (95% CI, 56% to 99%) for first booster (caution: only 12 boosted index cases during Delta). • Omicron: 45% (95% CI, −14% to 74%) for primary series; 64% (95% CI, 31% to 82%) for first booster. - Sensitivity analyses: • Including index cases with prior infection reduced VE-infectiousness: Delta 61% (95% CI, 3–84%) primary; 85% (95% CI, 10–98%) booster. Omicron 32% (95% CI, −11–59%) primary; 32% (95% CI, −5–57%) booster. • Time since vaccination: VE-infection and VE-infectiousness generally lower ≥3 months after vaccination. During Delta, VE-infection for primary series <3 months was 92% (95% CI, 66–98%) vs 12% (95% CI, −138–67%) ≥3 months. During Omicron, booster VE-infection was −3% (95% CI, −51–30%) <3 months vs −58% (95% CI, −138 to −5%) ≥3 months. • Shorter secondary-case window (7 days) produced estimates similar to main analysis. • Excluding household members <18 years increased VE-infectiousness and decreased VE-infection estimates. Overall, vaccines reduced infectiousness for Delta and Omicron; protection against infection was evident for Delta but not for Omicron in this household exposure context.

The study directly addresses the roles of COVID-19 vaccination in preventing infection and reducing onward transmission within households across Delta and Omicron periods. Findings show substantial reductions in infectiousness among vaccinated index cases, strongest for Delta and still meaningful for Omicron, supporting the concept that vaccination helps limit household transmission. VE-infection was moderate for Delta but not observed for Omicron, likely reflecting Omicron’s immune escape, intense household exposure, time since vaccination, and unmeasured prior infections among household members that may have reduced susceptibility independent of vaccination. Comparisons with other studies indicate broadly consistent reductions in infectiousness during Delta and heterogeneity for Omicron (especially early BA.1-focused studies). Including index cases with prior infection attenuated VE-infectiousness, suggesting natural infection confers some reduction in infectiousness and may dilute vaccine-attributable effects. Policy-wise, reductions in infectiousness support booster vaccination for those with close contact with vulnerable populations (e.g., healthcare and elderly care workers) to mitigate transmission risk, particularly when highly transmissible variants circulate.

COVID-19 vaccination reduced infectiousness of SARS-CoV-2 in household settings during both Delta and Omicron periods and reduced infection risk during Delta. The ability to demonstrate VE against Omicron infection was limited, likely due to immune escape, waning, intense household exposure, and unmeasured prior infections. The results support booster vaccination strategies for individuals in frequent contact with vulnerable populations to curb transmission. Future research should incorporate comprehensive data on prior infections among contacts, behavioral responses post-index diagnosis, and variant-specific effects (including newer subvariants), and evaluate timing and composition of booster regimens to optimize protection against both infection and transmission.

- Potential differential testing behavior and timing between vaccinated and unvaccinated index cases/household members; quarantine rules varied over time by vaccination status, possibly biasing estimates. - Secondary cases were assumed to be infected by the index within a 2–14 day window; some infections may have originated outside the household, potentially underestimating VE-infectiousness (similar estimates with a 7-day window mitigate this concern). - No data on behavioral changes post-index diagnosis (e.g., distancing), which could differ by vaccination status and bias transmission estimates. - Reliance on self-reported data for household members (and for index case vaccination when not in registry), introducing potential recall/reporting bias; mitigated by short recall interval. - Inability to adjust VE-infection for prior infections among household members likely diluted observed VE-infection, especially during Omicron. - Small numbers in some subgroups (e.g., boosted index cases during Delta, n=12), leading to imprecise estimates. - Time since vaccination was often >3 months, contributing to waning and lower VE estimates.