ARTICLE https://doi.org/10.1038/s41467-022-29301-9 OPEN Comparative effectiveness and safety of homologous two-dose ChAdOx1 versus heterologous vaccination with ChAdOx1 and BNT162b2 Eduardo Hermosilla 1,2, Ermengol Coma 1, Junqing Xie 3, Shuo Feng 4, Carmen Cabezas5, Leonardo Méndez-Boo 1, Francesc Fina1, Elisabet Ballo1, Montserrat Martínez5, Manuel Medina-Peralta1, 1234567890():,; Josep Maria Argimon5 & Daniel Prieto-Alhambra 3,6 ✉ Small trials have suggested that heterologous vaccination with first-dose ChAdOx1 and second-dose BNT162b2 may generate a better immune response than homologous vacci- nation with two doses of ChAdOx1. In this cohort analysis, we use linked data from Catalonia (Spain), where those aged <60 who received a first dose of ChAdOx1 could choose between ChAdOx1 and BNT162b2 for their second dose. Comparable cohorts were obtained after exact-matching 14,325/17,849 (80.3%) people receiving heterologous vaccination to 14,325/149,386 (9.6%) receiving homologous vaccination by age, sex, region, and date of second dose. Of these, 464 (3.2%) in the heterologous and 694 (4.8%) in the homologous groups developed COVID-19 between 1st June 2021 and 5th December 2021. The resulting hazard ratio (95% confidence interval) is 0.66 [0.59–0.74], favouring heterologous vacci- nation. The two groups had similar testing rates and safety outcomes. Sensitivity and negative control outcome analyses confirm these findings. In conclusion, we demonstrate that a heterologous vaccination schedule with ChAdOx1 followed by BNT162b2 was more efficacious than and similarly safe to homologous vaccination with two doses of ChAdOx1. Most of the infections in our study occurred when Delta was the predominant SARS-CoV-2 variant in Spain. These data agree with previous phase 2 randomised trials. 1 Direcció assistencial d’Atenció Primària i a la Comunitat, Institut Català de la Salut (ICS), Generalitat de Catalunya, Barcelona, Spain. 2 Idiap Jordi Gol, Universitat Autonoma de Barcelona, Barcelona, Spain. 3 Centre for Statistics in Medicine, NDORMS, University of Oxford, Oxford, UK. 4 Oxford Vaccine Group, Department of Paediatrics, University of Oxford, Oxford, UK. 5 Public Health Secretariat, Department of Health, Generalitat de Catalunya, Barcelona, Spain. 6 Department of Medical Informatics, Erasmus University Medical Center, Rotterdam, Netherlands. ✉email: daniel.prietoalhambra@ndorms.ox.ac.uk NATURE COMMUNICATIONS | (2022)13:1639 | https://doi.org/10.1038/s41467-022-29301-9 | www.nature.com/naturecommunications 1 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-29301-9 T he rapid development of vaccines to prevent Coronavirus Figure 1 shows the inclusion/exclusion steps used to identify the Disease 2019 (COVID-19) has allowed remarkable pro- study participants. For primary analyses, 14,325/17,849 (80.3%) gress in the global fight against the SARS-CoV-2 pan- people in the heterologous group were matched to 14,325/149,386 demic. As of 27 October 2021, around half of the world’s (9.6%) in the homologous group vaccinated with their second population had received at least one dose of a COVID-19 vaccine. dose on the same date (±2 days). The resulting cohorts were While 15 vaccines have been approved for use by at least one comparable in terms of all observed demographics, comorbidity, authority, most countries have received more than four approved medicine use, area of residence and socio-economic status vaccines1. (Supplementary Fig. 1). Exact matching ensured the same average The Chimpanzee Adenovirus-vectored Oxford (ChAdOx1) (SD) age (42.2 (9.6) years) and proportion of female participants and BioNTech/Pfizer mRNA (BNT162b2) vaccines were among (62.5%) in the two groups. A comparable proportion lived in the the first approved by the European Medicines Agency for emer- most socio-economically deprived (15.8% heterologous vs 15.9% gency use in the European Union. Both vaccines were tested in homologous) and rural (19.4% heterologous vs 19.5% homo- large phase 3 randomised controlled trials and found to be highly logous) areas of the country (Table 1). effective against symptomatic SARS-CoV-2 infection when given Study participants received their second doses between 27 as two doses2–5. Follow-up studies have demonstrated their April 2021 and 8 October 2021. Second doses occurred on the clinical effectiveness against severe disease, including preventing same date for 8742 (61.0%) matched pairs, 1 day apart (before/ hospitalisations and mortality, overall6,7 and in previously under- after) for 4302 (30.0%) pairs and 2 days apart for the remaining researched populations8. 1281 (9.0%) pairs. Test rates were similar in the matched cohorts, Spanish guidelines initially recommended the use of ChAdOx1 with 4874 (34.0%) people on the heterologous schedule and 3268 for people aged younger than 60 years due to the under- (35.8%) on the homologous schedule tested at least once during representation of elderly people in the initial pivotal trials2. Key the study period. The average (SD) number of tests for hetero- workers were targeted in this initial stage to maximise the impact logous and homologous groups was 0.82 (1.65) vs 0.87 (1.68) of vaccination on community transmission9. Despite their effi- overall and 2.42 (2.04) vs 2.44 (2.03) among those tested at least cacy, reports of thrombotic events after the first dose of once during follow-up, respectively. Table 2 shows number, type adenovirus-based COVID-19 vaccines led to recommendations of test (lateral flow test (LFT) vs polymerase chain reaction (PCR) for heterologous vaccination for those vaccinated with a first dose test) and test incidence rates according to vaccination regimen. of ChAdOx1, i.e. many European authorities recommended the The incidence rate ratio [95% confidence interval] of testing use of BNT162b2 for second doses to avoid further exposure to for heterologous versus homologous vaccination was 1.01 ChAdOx1. A small randomised controlled trial was rapidly [0.96–1.06]. Figure 2 shows that the matched cohorts had similar conducted that demonstrated better immunogenicity from het- timings for vaccination and testing over time. erologous vaccination in Spain10, but larger studies on clinical Between 1 June and 5 December 2021, SARS-CoV-2 infections effectiveness and safety are urgently needed11. were recorded for 464 (3.2%) people in the heterologous group, The Spanish authorities allowed citizens previously vaccinated equivalent to an incidence rate of 0.18/1000 person-years, and with a first dose of ChAdOx1 to choose between ChAdOx1 and 694 (4.8%) people in the homologous group, equivalent to an BNT162b2 for their second dose. The majority chose homologous incidence rate of 0.27/1000 person-years. These rates are vaccination with two doses of ChAdOx112. In the absence of equivalent to a hazard ratio of 0.66 [0.59–0.74], favouring het- phase 3/4 randomised controlled trials, this created a natural erologous vaccination (Fig. 3 and Table 2). This was equivalent to experiment for studying the comparative safety and effectiveness an absolute risk reduction (ARR) of 0.016 [0.012–0.021] in the of these two vaccination schedules. We leveraged routinely col- study period. Only 2 (0.01%) hospital admissions with COVID- lected health data, including electronic medical records linked to 19 were identified in the heterologous group, compared with 5 vaccination data and laboratory tests, to study the comparative (0.03%) in the homologous. No deaths were seen in either group. effectiveness and safety of homologous (two-dose ChAdOx1) and A total of 489 (3.4%) and 318 (2.2%) of the participants in the heterologous (ChadOx1 followed by BNT162b2) vaccination. heterologous and homologous cohorts received a third dose of an mRNA vaccine during follow-up, respectively, and were censored on that date. Results Regarding safety, primary analyses of 14,325 people per group Of 167,235 eligible people, 17,849 (10.7%) chose heterologous found only one venous thromboembolism event (0.007%) and vaccination and 149,386 (89.3%) chose homologous vaccination. one venous thromboembolism with thrombocytopenia event Fig. 1 Population flowchart. 2 NATURE COMMUNICATIONS | (2022)13:1639 | https://doi.org/10.1038/s41467-022-29301-9 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-29301-9 ARTICLE Table 1 Baseline characteristics of study participants CoV-2 infection were comparable to those in the primary ana- according to vaccination schedule. lysis: 0.65 [0.58–0.73] in the 1:2 and 0.65 [0.57–0.73] in the 1:5 matched cohorts. No safety concerns were identified in these larger cohorts, with only one additional safety event of myoper- Variable Heterologous Homologous icarditis identified in the homologous vaccination group for 1:5 N 14,325 14,325 matching (Table 3). Socio-demographic and socio-economic Finally, a post-hoc analysis was conducted using propensity Mean (SD) age, years 42.20 (9.60) 42.21 (9.57) scores and vaccine date matching instead of the matching Female sex 8959 (62.5%) 8959 (62.5%) Socio-economic status: first 2725 (19.02%) 2745 (19.16%) described above (see Methods). Comparable cohorts were quartile (least deprived) obtained after matching 17,173/17,849 (96.0%) of those vacci- Socio-economic status: second 4403 (30.74%) 4363 (30.46%) nated with heterologous regimen to 17,173/149,386 (11.5%) in quartile the homologous (Supplementary Fig. 3). A total of 551/17,173 Socio-economic status: third 2162 (15.09%) 2145 (14.97%) (3.2%) and 809/17,173 (4.7%) COVID-19 cases were identified quartile amongst those in the heterologous and homologous vaccination Socio-economic status: fourth 2260 (15.78%) 2276 (15.89%) cohorts, respectively, equivalent to incidence rates of 0.18/1000 quartile (most deprived) and 0.27/1000 person-years. The resulting HR was similar to that Residence in a rural area 2775 (19.37%) 2796 (19.52%) seen in the main analysis: HR 0.67 [0.60–0.75]. Medicines use Analgesics 654 (4.57%) 533 (3.72%) Sedatives/hypnotics 1049 (7.32%) 948 (6.62%) Discussion Anticoagulants 201 (1.40%) 116 (0.81%) Antidepressants 1228 (8.57%) 1077 (7.52%) To the best of our knowledge, this is the first report to date Antiepileptics 449 (3.13%) 353 (2.46%) comparing the safety and effectiveness of homologous vaccina- Antipsychotics 281 (1.96%) 157 (1.10%) tion against COVID-19 with two-dose ChAdOx1 and hetero- Antacids 618 (4.31%) 533 (3.72%) logous vaccination with first-dose ChAdOx1 and second-dose Systemic corticosteroids 101 (0.71%) 82 (0.57%) BNT162b2. Our primary analysis included over 28,000 people, Oral antidiabetics 187 (1.31%) 150 (1.05%) over 14,000 per group, exactly matched on age, sex, region and Insulin 105 (0.73%) 73 (0.51%) date of second-dose vaccination. In this rich linked cohort, we Lipid modifying agents 453 (3.16%) 410 (2.86%) found a 34% relative risk reduction of SARS-CoV-2 infection Alpha blockers 5 (0.03%) 2 (0.01%) (primary outcome) among those on the heterologous vaccination Other antihypertensives 4 (0.03%) 1 (0.01%) schedule compared with those on the homologous vaccination Beta blockers 202 (1.41%) 193 (1.35%) schedule, despite similar testing rates in the two groups. Calcium channel blockers 135 (0.94%) 101 (0.71%) No safety concerns were identified, with only one event Combination antihypertensives 209 (1.46%) 176 (1.23%) (<0.01%) of venous thromboembolism and one event of venous Diuretics 103 (0.72%) 108 (0.75%) ACE inhibitors/ARBs 423 (2.95%) 424 (2.96%) thromboembolism with thrombocytopenia in the heterologous Chronic obstructive pulmonary 579 (4.04%) 528 (3.69%) group in the main analysis. One additional event of myoper- disease/asthma inhalers icarditis was observed in 42,845 people receiving homologous Comorbidities vaccination in the 1:5 matched sensitivity analysis. Atrial fibrillation 23 (0.16%) 19 (0.13%) Sensitivity analyses using 1:2 and 1:5 matching increased the Osteoarthritis 513 (3.58%) 533 (3.72%) sample size to >37,000 and >50,000 participants, respectively, and Asthma 935 (6.53%) 910 (6.35%) confirmed the safety and effectiveness findings. Similarly, a post- Ischaemic heart disease 48 (0.34%) 38 (0.27%) hoc analysis using propensity score matching methods yielded Diabetes mellitus 266 (1.86%) 209 (1.46%) consistent findings. Liver disease 289 (2.02%) 278 (1.94%) The null association between vaccination schedule and our Hypertension 826 (5.77%) 814 (5.68%) chosen negative control outcome of back pain supported the Heart failure 5 (0.03%) 2 (0.01%) robustness of our findings, ruling out residual confounding. Cerebrovascular disease 41 (0.29%) 26 (0.18%) Chronic obstructive pulmonary 46 (0.32%) 37 (0.26%) Our findings that heterologous vaccination was more effective disease than homologous vaccination against COVID-19 agrees with Chronic kidney disease 44 (0.31%) 54 (0.38%) emerging efficacy evidence based on immunological endpoints. Cancer (all except non-melanoma 344 (2.40%) 325 (2.27%) Two small, randomised trials have reported higher immuno- skin cancer) genicity, characterised by humoral and cellular responses, Obesity 1539 (10.74%) 1391 (9.71%) from ChAdOx1/BNT162b2 than ChAdOx1/ChAdOx110,13. Valvular disease 73 (0.51%) 64 (0.45%) These results were also corroborated by several cohort Hepatitis B 20 (0.14%) 15 (0.10%) studies14–16. Although people vaccinated with heterologous Hepatitis C 49 (0.34%) 34 (0.24%) ChAdOx1/mRNA vaccine (e.g. BNT162b2) were reported to have HIV infection 49 (0.34%) 47 (0.33%) a 68% lower risk of symptomatic COVID-19 infection than unvaccinated people17, little was known about the comparative effectiveness of the heterologous and homologous vaccination (0.007%), both in the heterologous group. No myopericarditis schedules against clinical endpoints. Our study showed that events were seen in either group (Table 3). heterologous ChAdOx1/BNT162b2 vaccination conferred 34% Back pain episodes were used as a negative control outcome. more protection against SARS-CoV-2 infection than homologous They were recorded at similar frequencies in the two groups two-dose ChAdOx1 vaccination, and corroborate the potential of (Supplementary Fig. 2). immunological surrogate endpoints of COVID-19 vaccines being Sensitivity analyses were conducted using 1:2 and 1:5 exact predictive for clinical protection18,19. matching, resulting in 12,512 people with heterologous vaccina- Data on the post-marketing safety of heterologous vaccination tion matched to 25,024 people with homologous vaccination and schedules remain sparse, particularly for rare safety events, with 8569 matched to 42,845, respectively. The hazard ratios for SARS- most evidence from evaluations of reactogenicity10,13–15,20. NATURE COMMUNICATIONS | (2022)13:1639 | https://doi.org/10.1038/s41467-022-29301-9 | www.nature.com/naturecommunications 3 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-29301-9 Table 2 Number and incidence rate (per 1000 person-years) of tests and SARS-CoV-2 infection (positive test) following second- dose vaccination according to vaccination schedule. Heterologous Homologous HR/IRR N / Mean IR / SD N / Mean IR / SD [95% CI] Tested 4874 2.21/1000 py 5134 2.35/1000 py n/a Number of tests (overall) 0.82 1.65 0.87 1.68 1.01 [0.96–1.06] N of tests among the tested 2.42 2.04 2.44 2.03 n/a N of PCR tests (overall) 0.37 1.15 0.38 1.20 1.02 [0.96–1.08] N of PCR tests among the tested 1.86 1.98 1.87 2.08 n/a N of LFT tests overall 0.45 1.03 0.50 1.05 1.00 [0.96–1.04] N of LFT tests among the tested 2.18 1.17 2.17 1.07 n/a SARS-CoV-2 infection 464 0.18/1000 py 694 0.27/1000 py 0.66 [0.59–0.74] CI confidence interval, HR hazard ratio, IR incidence rate, IRR incidence rate ratio, N number, py person-years of follow-up, SD standard deviation. Fig. 2 Vaccine uptake and testing rates according to vaccination schedule. Fig. 3 Kaplan–Meier plot of COVID-19 infection (primary outcome) after second-dose vaccination according to vaccination schedule. Although reactogenicity endpoints are informative for assessing years23. In this study with >28,000 participants, we identified one potential vaccine side effects, these trials are underpowered to venous thromboembolism event and one venous thromboembo- study rare safety outcomes. Adverse events related to the ChA- lism with thrombocytopenia event. The number of events did not dOx1 vaccine include the rare (<1/1000 to ≥1/10,000) outcome increase much in sensitivity analyses including up to >50,000 venous thrombosis and the very rare (<1/10,000) outcome participants, suggesting that larger studies are needed to investi- vaccine-induced immune thrombosis with thrombocytopenia gate these safety signals for heterologous vaccination schedules. syndrome21,22. Similarly, myocarditis and pericarditis outcomes Our study has several limitations. The main limitation is the possibly associated with BNT162b2 are expected to affect around observational nature of our data. However, exact matching on 10–24 people per 10 million fully vaccinated people aged ≥30 age, region, and date led to a good balance in all observed 4 NATURE COMMUNICATIONS | (2022)13:1639 | https://doi.org/10.1038/s41467-022-29301-9 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-29301-9 ARTICLE Table 3 Number (%) of safety events in the 21 days following second dose, according to vaccination schedule. Heterologous Homologous 1:1 matching 1:2 matching 1:5 matching 1:1 matching 1:2 matching 1:5 matching N participants 14,325 12,512 8569 14,325 25,024 42,845 N (%) of venous thromboembolism events 1 1 1 0 0 0 N (%) of venous thromboembolism with 1 1 1 0 0 0 thrombocytopenia N (%) of myopericarditis 0 0 0 0 0 1 confounders, including socio-demographics, comorbidity and and a second dose of ChAdOx1 (homologous vaccination) or BNT162b2 (het- medicines use. Propensity score matching was used in a post-hoc erologous vaccination). We followed participants from the day they received their second dose of either vaccine until an outcome, death, third dose of the vaccine, or analysis to confirm the robustness of our exact-matching the end of data availability (5 December 2021). approach. Additionally, our analysis of a negative control out- We excluded people with a previous SARS-CoV-2 infection identified by a come (back pain) suggested comparability of the matched positive RT-PCR test or LFT and people assigned to one of the 10% of primary- cohorts, including unobserved covariates. care practices not contributing to our database. As most of our participants were middle-aged adults aged less Each participant receiving heterologous vaccination was matched 1:1 to one person receiving homologous vaccination using exact matching by age, sex, general than 60 years old, our risk-benefit assessment may not be valid practice centre and date of second dose. In a sensitivity analysis, we changed the for younger or elderly people. In addition, we could not include matching ratio to 1:2 and 1:5 to increase sample size. the cohort of people receiving 2 doses of BNT162b2 in our analyses, as they were vaccinated during a different calendar Study outcomes. The primary outcome for effectiveness analyses was SARS-CoV-2 period, as depicted in Supplementary Fig. 4. Failing to match on infection, defined by the date of the earliest of a positive RT-PCR test or LFT, vaccination date would lead to biased results due to changes in regardless of symptoms or clinical diagnosis. We measured the number of tests over time regardless of results as an additional outcome to account for diagnostic effort. community transmission during the study period. Finally, our Safety outcomes included venous thromboembolism, venous thromboembolism sample size was insufficient for studying severe COVID-19 out- with thrombocytopenia and myopericarditis within 21 days after the second comes, including hospitalisation and mortality, or rare safety vaccine dose, based on ChAdOx121 and BNT162b223 safety reports. outcomes. Supplementary Table S2 includes the ICD-10-CM codes (international This study also has strengths. The rich, representative linked classification of diseases, 10th revision, clinical modification) used to ascertain when these events occurred. dataset used allowed a robust analysis of vaccine exposure and We analysed the occurrence of a negative control outcome—low back pain—to outcomes at speed to inform ongoing international vaccination identify potential unmeasured confounding. Negative control outcomes are health campaigns. Catalonia has a universal healthcare system and uses events not causally associated with the exposure of interest, here vaccination. a centralised, secure data ecosystem with a long track record of research and high-impact publications8,24. The granularity of Additional covariates. Covariates used for confounding assessment included these data made it possible to control for confounding and test socio-demographics and clinical features assessed at the time of inclusion (day of for residual systematic bias. Linkage to additional data sources the second vaccination), as recorded in primary-care electronic health records and linked administrative data: age (in years), sex, area of residence, rurality and socio- on RT-PCR and LFT tests allowed for a comprehensive assess- economic status, number of RT-PCR tests or LFT performed, pre-existing ment of testing rates and reliable SARS-CoV-2 infection rates. comorbidities and long-term medicine use. Supplementary Table 1 provides ICD- The data collection period covered a time when most cases of 10-CM codes for comorbidities and Anatomical Therapeutic Chemical Classifi- COVID-19 in Catalonia were attributable to the Delta variant of cation (ATCC) codes used to identify previous medicine use. We assessed socio- economic status using a validated deprivation index based on census data (MEDEA SARS-CoV-225,which is still the predominant variant worldwide. deprivation index)26,27. Rurality of residence was measured, with rural areas Our data are highly relevant for ongoing global vaccination defined by a population <10,000 inhabitants and a density <150 inhabitants/km2, strategies and future and current third-dose and booster as per regional guidance. campaigns. In conclusion, by leveraging the potential of multiple data Statistical analysis. Exact matching (1:1) between heterologous ChAdOx1/ sources in parallel, our study confirmed that a heterologous BNT162b2 vaccination and homologous two-dose ChAdOx1 was performed using vaccination schedule of ChAdOx1 and BNT162b2 was safe and the following variables: age (±2 years), sex, general practice centre and day of the second vaccination (±2 days). As a sensitivity analysis, we generated additional provided better protection against COVID-19 than a homologous study populations and repeated all analyses after matching with 1:2 and 1:5 ratios. ChAdOx1 vaccination schedule in real-world settings experien- We assessed confounding due to known variables by measuring covariate imbal- cing the Delta variant. More research on other mixed-vaccine ance as the standardised mean difference (SMD) of all covariates listed above. We schedules with different prime-boost intervals are needed. considered SMD > 0.1 to be imbalanced28. Separately, we used propensity score matching as part of a post-hoc analysis requested by one of the manuscript reviewers. Propensity scores were estimated Methods using logistic regression including all available confounders, and cohorts matched Study design and data sources. We performed a cohort study based on linked using propensity score values with a maximum caliper width of 0.2 SD, with routinely collected data available to the Public Health Secretariat of Catalonia. additional exact matching on second-dose vaccination date (±2 days). Vaccine exposure was obtained from the Catalan Shared Clinical Records, a We plotted time-to-event Kaplan–Meier estimates according to vaccine database with vaccine data covering the entire Catalan health system and all its exposure (homologous vs heterologous). Absolute risk reduction (ARR) was vaccination centres. Additional linked data were obtained from the Catalan data- estimated as the difference in cumulative incidence of Covid-19 amongst those base of reverse transcription polymerase chain reaction (RT-PCR) tests and lateral receiving homologous–heterologous vaccination. Cox regression models were then flow tests (LFT) for SARS-CoV-2, from primary-care electronic health records and fitted to calculate hazard ratios and 95% confidence intervals for each of the study from a population-based administrative hospital admissions data (CMBD-AH for outcomes, according to vaccination schedule. Visual inspection of Schoenfeld its acronym in Catalan language). Data from these linked databases have previously residuals against the transformed time was used to evaluate the proportionality of been used for multiple COVID-19 research studies and include information for hazards. A zero-inflated negative binomial regression model was used to calculate nearly 90% of the Catalan population8. the incident rate ratio and 95% confidence interval for the number of tests. All analyses were conducted using R version 4.0.0. Participants, cohorts and follow-up. For our primary analysis, we included all Ethical considerations and information governance. All data were obtained from individuals aged 19–59 years old who received a first dose of the ChAdOx1 vaccine linked administrative sources after pseudonymisation in accordance with articles 6. NATURE COMMUNICATIONS | (2022)13:1639 | https://doi.org/10.1038/s41467-022-29301-9 | www.nature.com/naturecommunications 5 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-29301-9 e), 9.2. j) + 89 RGPD and 17.2.d of the LOPD-GDD. The proposed study was symptomatic Covid-19 infection in Sweden: a nationwide cohort study. Lancet evaluated and approved by the Clinical Research Ethics Committee of the IDIAP Reg. Health Euro https://doi.org/10.1016/j.lanepe.2021.100249 (2021). Jordi Gol with Reference 21/269-PCV. This research was based on the agreement 18. Jin, P., Li, J., Pan, H., Wu, Y. & Zhu, F. Immunological surrogate endpoints of established in Regulation 2016/679 of the European Parliament and of the Council COVID-2019 vaccines: the evidence we have versus the evidence we need. of April 27, 2016 on Data Protection and Organic Law 3/2018 of December 5 of Signal Transduct. Target. Ther. 6, 48 (2021). protection of personal data and guarantee of digital rights. 19. Feng, S. et al. Correlates of protection against symptomatic and asymptomatic SARS-CoV-2 infection. Nat. Med. https://doi.org/10.1038/s41591-021-01540-1 Reporting summary. Further information on research design is available in the Nature (2021). Research Reporting Summary linked to this article. 20. Powell, A. A. et al. Real-world data shows increased reactogenicity in adults after heterologous compared to homologous prime-boost COVID-19 vaccination, March-June 2021, England. Eurosurveillance 26, 2100634 (2021). Data availability 21. Li, X. et al. Characterising the background incidence rates of adverse events of Patient-level data utilised for the reported analyses are not accessible openly due to local special interest for covid-19 vaccines in eight countries: multinational network information governance regulations as defined in the ‘Ethical considerations and cohort study. BMJ 373, n1435 (2021). information governance’ section. Aggregated data are provided for inspection as 22. Council for International Organizations of Medical Sciences. Guidelines for Supplementary Data 1-3, and R code used for analyses as Supplementary Code 1. Preparing Core Clinical-Safety Information on Drugs, Second Edition, Report of CIOMS Working Group III and V. https://cioms.ch/wp-content/uploads/2018/ Received: 12 January 2022; Accepted: 9 March 2022; 03/Guidelines-for-Preparing-Core-Clinical-Safety-Info-Drugs-Report-of- CIOMS-Working-Group-III-and-V.pdf (2001). 23. Gargano, J. W. et al. Use of mRNA COVID-19 vaccine after reports of myocarditis among vaccine recipients: update from the Advisory Committee on Immunization Practices—United States, June 2021. MMWR Morb. Mortal. Wkly Rep. 70, 977–982 (2021). 24. Xie, J. et al. Association of tramadol vs codeine prescription dispensation References with mortality and other adverse clinical outcomes. JAMA 326, 1504–1515 1. UNICEF COVID-19 Vaccine Market Dashboard. https://www.unicef.org/ (2021). supply/covid-19-vaccine-market-dashboard. (2021). 25. Català, M. et al. Analysis of new variants of SARS-CoV-2 in the Barcelona city 2. Voysey, M. et al. Safety and efficacy of the ChAdOx1 nCoV-19 vaccine and Northern Metropolitan area: success of the Delta and sinking of the (AZD1222) against SARS-CoV-2: an interim analysis of four randomised B.1621 and the Gamma. in Analysis and prediction of COVID-19 for EU- controlled trials in Brazil, South Africa, and the UK. Lancet 397, 99–111 EFTA-UK and other countries, #260. Research Report, 10–14 (Universitat (2021). Politècnica de Catalunya. 2021). 3. Voysey, M. et al. Single-dose administration and the influence of the timing of 26. Domínguez-Berjón, M. F. et al. [Constructing a deprivation index based on the booster dose on immunogenicity and efficacy of ChAdOx1 nCoV-19 census data in large Spanish cities (the MEDEA project)]. Gac. Sanit. 22, (AZD1222) vaccine: a pooled analysis of four randomised trials. Lancet 397, 179–187 (2008). 881–891 (2021). 27. Garcia-Gil, M. et al. Linking of primary care records to census data to study 4. Voysey, M. & Pollard, A. J. ChAdOx1 nCoV-19 vaccine for SARS-CoV-2— the association between socioeconomic status and cancer incidence in Authors’ reply. Lancet 396, 1486–1487 (2020). Southern Europe: a nation-wide ecological study. PLoS ONE 9, e109706 5. Polack, F. P. et al. Safety and efficacy of the BNT162b2 mRNA Covid-19 (2014). vaccine. N. Engl. J. Med. 383, 2603–2615 (2020). 28. Nguyen, T. L. et al. Double-adjustment in propensity score matching analysis: 6. Sheikh, A., McMenamin, J., Taylor, B. & Robertson, C. SARS-CoV-2 Delta choosing a threshold for considering residual imbalance. BMC Med. Res. VOC in Scotland: demographics, risk of hospital admission, and vaccine Methodol. 17, 78 (2017). effectiveness. Lancet 397, 2461–2462 (2021). 7. Haas, E. J. et al. Impact and effectiveness of mRNA BNT162b2 vaccine against SARS-CoV-2 infections and COVID-19 cases, hospitalisations, and deaths Acknowledgements following a nationwide vaccination campaign in Israel: an observational study We acknowledge English language editing by Jennifer A de Beyer of the Centre for using national surveillance data. Lancet 397, 1819–1829 (2021). Statistics in Medicine, University of Oxford. We thank Luis Garcia-Eroles for assistance 8. Cabezas, C. et al. Associations of BNT162b2 vaccination with SARS-CoV-2 and support with data management regarding linked hospital admissions data for the infection and hospital admission and death with covid-19 in nursing homes study of safety outcomes. and healthcare workers in Catalonia: prospective cohort study. BMJ 374, n1868 (2021). 9. Grupo de Trabajo Técnico de Vacunación COVID-19, de la Ponencia de Author contributions Programa y Registro de Vacunaciones. Estrategia de vacunación frente a E.H., E.C., C.C. and D.P.A. designed the study. J.X., X.F., S.F., E.H., E.C. and D.P.A. COVID-19 en España. 9 de febrero de 2021. https://www.mscbs.gob.es/ drafted the first version of the manuscript. E.H., E.C., C.C., L.M.B., E.B., M.M., F.F., profesionales/saludPublica/prevPromocion/vacunaciones/covid19/docs/ J.M.A. and M.M.P. contributed to data collection, extraction and/or curation. E.H. and COVID-19_Actualizacion3_EstrategiaVacunacion.pdf. (2021). E.C. led data analyses. E.H., E.C., J.X., S.F., C.C., L.M.B., F.F., E.B., M.M., M.M.P., J.M.A. 10. Borobia, A. M. et al. Immunogenicity and reactogenicity of BNT162b2 booster and D.P.A. reviewed critically the content of the manuscript, edited further versions and in ChAdOx1-S-primed participants (CombiVacS): a multicentre, open-label, approved the final version and revisions for submission. randomised, controlled, phase 2 trial. Lancet 398, 121–130 (2021). 11. Duarte-Salles, T. & Prieto-Alhambra, D. Heterologous vaccine regimens Competing interests against COVID-19. Lancet 398, 94–95 (2021). D.P.A.’s research group has received research grants from the European Medicines 12. Grupo de Trabajo Técnico de Vacunación COVID-19, de la Ponencia de Agency, from the Innovative Medicines Initiative, from Amgen, Chiesi and from UCB Programa y Registro de Vacunaciones. Estrategia de vacunación frente a Biopharma; and consultancy or speaker fees from Astellas, Amgen, Astra Zeneca and COVID-19 en España. https://www.mscbs.gob.es/profesionales/saludPublica/ UCB Biopharma. The remaining authors declare no competing interests. prevPromocion/vacunaciones/covid19/docs/COVID-19_Actualizacion8_ EstrategiaVacunacion.pdf. (2021) 13. Liu, X. et al. Safety and immunogenicity of heterologous versus homologous Additional information prime-boost schedules with an adenoviral vectored and mRNA COVID-19 Supplementary information The online version contains supplementary material vaccine (Com-COV): a single-blind, randomised, non-inferiority trial. Lancet available at https://doi.org/10.1038/s41467-022-29301-9. 398, 856–869 (2021). 14. Schmidt, T. et al. Immunogenicity and reactogenicity of heterologous Correspondence and requests for materials should be addressed to Daniel Prieto- ChAdOx1 nCoV-19/mRNA vaccination. Nat. Med. 27, 1530–1535 (2021). Alhambra. 15. Hillus, D. et al. Safety, reactogenicity, and immunogenicity of homologous and heterologous prime-boost immunisation with ChAdOx1 nCoV-19 and Peer review information Nature Communications thanks the anonymous reviewers for BNT162b2: a prospective cohort study. Lancet Respir. Med. 9, 1255–1265 their contribution to the peer review of this work. Peer reviewer reports are available. (2021). 16. Tenbusch, M. et al. Heterologous prime-boost vaccination with ChAdOx1 Reprints and permission information is available at http://www.nature.com/reprints nCoV-19 and BNT162b2. Lancet Infect. Dis. 21, 1212–1213 (2021). Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in 17. Nordström, P., Ballin, M. & Nordström, A. Effectiveness of heterologous ChAdOx1 nCoV-19 and mRNA prime-boost vaccination against published maps and institutional affiliations. 6 NATURE COMMUNICATIONS | (2022)13:1639 | https://doi.org/10.1038/s41467-022-29301-9 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-29301-9 ARTICLE Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/ licenses/by/4.0/. © The Author(s) 2022 NATURE COMMUNICATIONS | (2022)13:1639 | https://doi.org/10.1038/s41467-022-29301-9 | www.nature.com/naturecommunications 7