key: cord-0264986-11q95qgf authors: Gram, M. A.; Emborg, H.-D.; Moustsen-Helms, I. R.; Nielsen, J.; Samposlashrensen, A. K. B.; Valentiner-Branth, P.; Schelde, A. B.; Nielsen, K. F. title: Vaccine effectiveness when combining the ChAdOx1 vaccine as the first dose with an mRNA COVID-19 vaccine as the second dose date: 2021-07-28 journal: nan DOI: 10.1101/2021.07.26.21261130 sha: 61af86a672b6ad973835f21293a152369fa60cc3 doc_id: 264986 cord_uid: 11q95qgf Background: The recommendations in several countries to stop using the ChAdOx1 vaccine has led to vaccine programs combining different vaccine types, which necessitates new knowledge on vaccine effectiveness (VE). In this study, we aimed to estimate the VE when combining the ChAdOx1 vaccine as the first dose and an mRNA vaccine as the second dose. Methods: This nationwide population-based cohort study estimated VE against SARS-CoV-2 infection, all-cause and COVID-19 related hospitalization and death after receiving the ChAdOx1 vaccine as the first dose followed by an mRNA vaccine as the second dose. VE estimates were obtained using a Cox regression with calendar time as underlying time and adjusted for sex, age, comorbidity, heritage and hospital admission. Information on all individuals was extracted and linked from high-quality national registries. Results: A total of 5,542,079 individuals were included in the analyses (97.6% of the total Danish population). A total of 144,360 were vaccinated with the ChAdOx1 vaccine as the first dose and of these 136,551 individuals received an mRNA vaccine as the second dose. A total of 1,691,464 person-years and 83,034 cases of SARS-CoV-2 infection were included. The VE against SARS-CoV-2 infection when combining the ChAdOx1 and an mRNA vaccine was 88% (95% confidence interval (CI): 83; 92) 14 days after the second dose and onwards. There were no COVID-19 related hospitalizations and deaths among the individuals vaccinated with the combination of the ChAdOx1 and an mRNA vaccine during the study period. Conclusion: In conclusion, this study found a reduction in the risk of SARS-CoV-2 infection when combining the ChAdOx1 and an mRNA vaccine, compared with unvaccinated individuals. This is similar to the VE of two doses of an mRNA vaccine. Longer follow-up time is needed to confirm vaccine induced protection against severe events, such as COVID-19 related hospitalization and death. similar to the VE of two doses of an mRNA vaccine. Longer follow-up time is needed to confirm vaccine induced protection against severe events, such as COVID-19 related hospitalization and death. (4). Studies from the UK have reported VE estimates between 22% and 94% following the administration of one dose of ChAdOx1 (5-7). Due to changing recommendations regarding the use of the ChAdOx1 vaccine (4) and to avoid vaccine shortages, some countries are combining vaccine types. This creates a need for studies of VE for heterologous vaccination schedules (8). Immunological data on a heterologous vaccination schedule indicates that the combination of the ChAdOx1/mRNA vaccines is at least as immunogenic and protective as homologous BNT162b2 vaccination (9, 10). However, no previous studies have reported VE of a ChAdOx1/mRNA vaccine schedule. This study aimed to estimate VE against SARS-CoV-2 infection, all-cause and COVID-19 related hospitalization and death of 1) one dose of the ChAdOx1 vaccine and 2) the ChAdOx1/mRNA vaccine schedule, compared with unvaccinated individuals. All residents in Denmark are registered in the Danish Civil Registration System and assigned a unique personal identification number (CPR number), which is used in all national registries, enabling accurate individual-level linkages between registries (11). In this nationwide retrospective population-based cohort study, individuals were included in the study population if they were residents in Denmark on 9 th of February 2021 or immigrated before the end of study on 23 rd of June 2021. Individuals who had received a COVID-19 vaccine or had a Reverse Transcription Polymerase Chain Reaction (RT-PCR) confirmed SARS-CoV-2 infection before the start of the study (2.4%) were excluded. The latter due to expected natural immunity from previousSARS-CoV-2 infection (12). The . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 28, 2021. ; https://doi.org/10.1101/2021.07.26.21261130 doi: medRxiv preprint study participants were followed from 9 th of February 2021 and until a SARS-CoV-2 infection, receiving the first dose of another COVID-19 vaccine than the ChAdOx1 vaccine, emigration, death or end of follow-up (23 rd of June 2021), whichever came first. Information on immigration, emigration and vital status was retrieved from the Danish Civil Registration System (11). Information on the date of vaccination and type of vaccine was retrieved from the Danish Vaccination Registry, which includes data on all administered vaccines (13) . The general exposures of interest were: 1) a single dose of ChAdOx1 and 2) a second dose of either BNT162b2 mRNA or mRNA-1273 following the first dose of ChAdOx1. Unvaccinated individuals were used as reference. To examine the effect of one dose of ChAdOx1 on SARS-CoV-2 infection, time following vaccination was divided into 0-13 days (the run-in period), and 14 days and onward was divided into 7-day-intervals until receiving the second dose. The effect of one dose of ChAdOx1 on hospitalization and death was divided into 0-13 days (the run-in period) and 14 days and onwards days until receiving the second dose of an mRNA vaccine. The effect of the ChAdOx1/mRNA vaccine schedule on all outcomes was divided into 0-13 days and 14 days and onwards. The outcomes of interest were a SARS-CoV-2 infection, defined as a laboratory confirmed RT-PCR SARS-CoV-2 positive test, and all-cause or COVID-19 related hospitalization and death. Information on RT-PCR SARS-CoV-2 positive tests was retrieved from the Danish Microbiology Database (MiBa), . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 28, 2021. ; https://doi.org/10.1101/2021.07.26.21261130 doi: medRxiv preprint which is a national database that automatically accumulates both positive and negative test results from all Danish departments of clinical microbiology (14) . Information about rapid antigen tests was not included due to moderate sensitivity in asymptomatic patients compared with RT-PCR (22). A COVID-19 related hospitalization or death was defined as an admission within 14 days or death within 30 days after a positive SARS-CoV-2 test, respectively. Therefore, time at risk for the COVID-19 related outcomes were extended to 14 and 30 days after a confirmed SARS-CoV-2 infection, respectively. Hospital admissions and discharge dates were retrieved from the Danish National Patient Registry (15) . The incidence of SARS-CoV-2 varied considerably throughout the study period. To control for this variation through the study period, calendar time was used as underlying time. Further, age, sex, heritage and comorbidity were also included as covariates for the association between COVID-19 vaccination and SARS-CoV-2 infection, hospitalization and death. Finally, hospital admission was included as a covariate for the association between COVID-19 vaccination and SARS-CoV-2 infection and death. Information on date of birth and sex (male/female) was retrieved from the Danish Civil Registration System (11). The presence of comorbidity (yes/no) within the previous five years (data retrieved at start of study) was identified based on diagnoses coded according to the International Classification of Diseases, 10th revision (ICD-10). Diagnoses included one primary and optional secondary diagnosis for each patient contact and were retrieved from the Danish National Patient Registry (15) . The ICD-10 codes included in comorbidity are provided in the supporting information (Table S1 ). Information on heritage (Danish/western/non-western/unknown) was retrieved from . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 28, 2021. Characteristics of the included population were described using proportions. Crude VE estimates were obtained using a Poisson regression with overdispersion and time at risk as offset. Adjusted The study population included 5,542,079 individuals, among which 144,360 (2.6%) received the ChAdOx1 vaccine as the first dose. Of these, 88,050 (61%) and 48,501 (33.6%) received the BNT162b2 mRNA and the mRNA-1273 vaccine as the second dose, respectively. A total of 1,691,464 person-years and 83,034 SARS-CoV-2 infections were included. The study population had a median age of 45 and 46 years at the time of the first and second dose, respectively. There was an equal distribution of males and females. Comorbidity was present among 25.6% of the individuals and the majority had Danish heritage (84%) ( Table 1) . . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 28, 2021. Significant adjusted VE estimates against SARS-CoV-2 infection between 14-83 days after one dose of the ChAdOx1 vaccine were relatively stable with overlapping CIs. However, the VE estimates were not significant at 0-13 days, 42-48 days and from 84 days (Fig 2) . The significant adjusted VE estimates ranged from 29% (95% CI: 12; 43) to 44% (95% CI: 32; 55) after one dose of the ChAdOx1 vaccine. For the ChAdOx1/mRNA vaccine schedule, significant adjusted VE estimates of 66% (95% CI: 59; 72) and 88% (95% CI: 83; 92) were observed at 0-13 days and from 14 days and onwards after the second dose, respectively (Table 2) . For all-cause hospitalization, significant adjusted VE estimates were observed at both 0-13 days (VE=36%, 95% CI: 28; 42) and from 14 days (VE=36%, 95% CI: 33; 40) after the first dose of ChAdOx1 and until receiving a second dose of an mRNA vaccine, respectively. Adjusted VE estimates of 43% . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 28, 2021. against COVID-19 related hospitalization was observed from 14 days after the first dose and until receiving a second dose of an mRNA vaccine. No COVID-19 related hospitalizations occurred after the ChAdOx1/mRNA vaccine schedule. Therefore, it was not possible to estimate VE estimates against COVID-19 related hospitalization for the ChAdOx1/mRNA vaccine schedule (Table 3) . For all-cause death, a significant adjusted VE of 48% (95% CI: 23; 65) was observed at 14 days after one dose of the ChAdOx1 vaccine and until receiving a second dose of an mRNA vaccine. The VE estimates increased to 93% (95% CI: 53; 99) at 0-13 days and 92% (95% CI: 69; 98) at 14 days onwards after an mRNA vaccine as the second dose. No COVID-19 related deaths occurred after receiving a COVID-19 vaccination and therefore a VE could not be estimated (Table 4 ). This nationwide population-based cohort study showed a significant reduction in the risk of SARS- This was expected due to the run-in period before immunity is anticipated to occur. Previous studies have shown similar results of one dose of the ChAdOx1 vaccine (5-7). A test-negative case-control study from England including adults aged 70 years and older showed a VE of 22% (95% CI: 11; 32) from 14-20 days, reaching a VE of 73% (95% CI: 27; 90) at 35 days and onwards after one dose of the ChAdOx1 vaccine (5). Additional protection against hospital admission was observed, suggesting a VE against emergency hospital admission of 80% (5). A cohort study from England, including longterm care facility residents aged 65 years and older, showed adjusted HR against infection . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 28, 2021. The strengths of this study are the high-quality registers and the possibility to use the unique personal identifier to link data on all residents in Denmark. The national testing strategy during the study period, including unlimited access to free-of-charge RT-PCR tests nationwide, led to a high proportion of the population being tested, which enabled us to capture data on both asymptomatic and symptomatic infections. Another strength was the access to national data on all laboratory confirmed RT-PCR SARS-CoV-2 infections. Also, a high sensitivity (97.1%) and specificity (99.98%) . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 28, 2021. ; https://doi.org/10.1101/2021.07.26.21261130 doi: medRxiv preprint was observed for the RT-PCR test (21) ensuring a low risk of misclassification. An effort was made to ensure that all individuals had equal opportunities to receive the COVID-19 vaccines. This was done through an online booking system, special campaigns, offering vaccination in some workplaces, translating the material to other languages than Danish and English and arranging transport for those who were not able to reach the vaccination clinics on their own. This study has also several limitations. Since we were not able to discriminate between asymptomatic and symptomatic infections, it was not possible to assess the severity of a COVID-19 infection after vaccination based on symptoms. We used a positive SARS-CoV-2 test prior to hospitalization and death (i.e. COVID-19 related hospitalization and death) as proxy for the severity of COVID-19, although this definition might be subject to uncertainty since COVID-19 may not be the cause for these outcomes. Another limitation is that differences in test behavior related to vaccination status may exist, this could result in capturing less asymptomatic infections in vaccinated individuals and lead to elevated VE estimates. The Cox regression models were adjusted for potential confounders including calendar time, age, sex, heritage, hospital admission and comorbidity. However, we cannot eliminate differences in health-seeking behavior or test activity and residual confounding from comorbidity being classified as a dichotomous covariate. This classification does not allow for the exclusion of differences in individual comorbidities across vaccination status, such as the vaccinated cohort being more or less severely ill than the unvaccinated cohort. It was mainly front-line personnel who received the ChAdOx1 vaccine (99.3%). Therefore, we cannot eliminate confounding by indication, assuming that front-line personnel are more exposed to SARS-CoV-2 than the general population (23, 24), which could result in a lower VE in this group. However, front-line personnel are also better trained to use personal protective equipment than the general population, which would moderate the increased risk from their . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 28, 2021. ; https://doi.org/10.1101/2021.07.26.21261130 doi: medRxiv preprint occupational setting. Through the study period, many individuals received COVID-19 vaccines other than ChAdOx1/mRNA vaccines, thus increasing the overall vaccine coverage in the general population, and thereby creating indirect protective herd immunity. In conclusion, VE against SARS-CoV-2 infection was 88% of the ChAdOx1/mRNA vaccine schedule which is similar to the VE of two doses of the BNT162b2 mRNA vaccine. A single dose of the ChAdOx1 vaccine seems to be protective against SARS-CoV-2 infection after 14 days and up to 83 days, but the second dose of an mRNA vaccine is needed to maintain a significant VE. No COVID-19 related hospitalizations were observed after the second dose, and no COVID-19 related deaths were observed after neither the first dose ChAdOx1 nor the ChAdOx1/mRNA vaccine schedule. This research did not receive any funding. All authors declared no competing interest. . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 28, 2021. . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 28, 2021. . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 28, 2021. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 28, 2021. ; https://doi.org/10.1101/2021.07.26.21261130 doi: medRxiv preprint Coronavirus disease (COVID-19) pandemic: World Health Organization The Danish vaccination register Real-time surveillance of laboratory confirmed influenza based on the Danish microbiology database (MiBa) The Danish National Patient Registry: a review of content, data quality, and research potential Moustsen-Helms Fig 2. Unadjusted and adjusted VE estimates against RT-PCR SARS-CoV-2 infection of one dose of the ChAdOx1 vaccine and the ChAdOx1/mRNA vaccine schedule The authors are grateful to the Danish Health Data Authority for their help in defining the population. We would also like to thank the Department of Data Integration and Analysis at Statens Serum Institut for data management.