Dear Editor, Limited supply of different vaccines at different timeframes has compelled the use of heterologous schedules in some countries. In these heterologous schedules, individuals were allocated a mixture of COVID-19 vaccine products. This strategy has been used in settings where the availability of follow-up with a similar vaccine may not be guaranteed. Studies of heterologous vaccine schedules have shown that a combination of different vaccines are safe and may increase the vaccine effectiveness for symptomatic and hospitalized Omicron COVID-19 variants and can be recommended in the population 1Au W.Y. Cheung P.P.H. Effectiveness of heterologous and homologous covid-19 vaccine regimens: living systematic review with network meta-analysis.BMJ. 2022; 377https://doi.org/10.1136/BMJ-2022-069989Crossref Google Scholar. In Colombia, the health authorities encouraged heterologous vaccination given the lack of specific vaccine products to complete homologous vaccination schedules2Ministry of Health . Vacunación contra COVID-19. 2023 . 〈https://www.minsalud.gov.co/salud/publica/Vacunacion/Paginas/Vacunacion-covid-19.aspx〉 (accessed April 2, 2023).Google Scholar. Thus, we aimed to estimate the vaccine effectiveness of heterologous mRNA and non-mRNA booster vaccination schedules of five vaccines for the Omicron variant of COVID-19. We conducted a test-negative case-control study in the adult population of Bogota (Colombia). We selected cases of symptomatic laboratory confirmed COVID-19 (using rapid or polymerase-chain-reaction tests), in time-periods of Omicron predominance (>99% of samples, SFig. 1). Test-negative controls were selected using negative polymerase-chain-reaction tests. Adjusted logistic regression analyses were used to estimate vaccine effectiveness (one minus the odds ratio), with 95% confidence intervals (CI). The Supplementary Material details our methods. A total of 159,088 patients were ascertained for inclusion during Omicron (from January 7th to February 24th of 2022). After the removal of missing data, we analyzed a total of 43,200 cases (SFigure 2). The Table 1 describes the differences between cases and controls. The adjusted effectiveness of mRNA primary series with booster mRNA vaccines was larger than schedules with mRNA primary series plus non-mRNA boosters (P<0.001) and non-mRNA primary series with non-mRNA booster vaccine doses (P<0.001) (Panel A of the Fig. 1). Similarly, mRNA booster schedules with non-mRNA primary series were more effective than non-mRNA primary series with non-mRNA boosters (P<0.001). The schedules consisting of only mRNA vaccines were the most effective for the prevention of Omicron, and were more effective than non-mRNA primary series with booster mRNA (P=0.027).Table 1Characteristics of symptomatic cases of laboratory confirmed positive Omicron patients and test-negative controls, in Bogota, Colombia.CharacteristicsControlsOmicron casesP-valuen = 7,000 (%)n = 43,200 (%)Age (years), median (25th-75th percentiles)51.0 (34.0-66.0)43.0 (31.0-60.0)<0.001Age (years)<0.001 18-392,455 (35.1)18,901 (43.8) 40-591,967 (28.1)13,109 (30.3) 60-792,013 (28.8)9,254 (21.4) 80+565 (8.1)1,936 (4.5)Sex<0.001 Female4,306 (61.5)24,217 (56.1) Male2,694 (38.5)18,983 (43.9)Asthma274 (3.9)822 (1.9)<0.001COPD310 (4.4)608 (1.4)<0.001Diabetes483 (6.9)1,590 (3.7)<0.001HIV54 (0.8)98 (0.2)<0.001Cardiac disease335 (4.8)1,002 (2.4)<0.001Cancer209 (3.0)589 (1.4)<0.001Malnutrition37 (0.5)49 (0.1)<0.001Obesity371 (5.3)973 (2.3)<0.001Renal failure106 (1.5)243 (0.6)<0.001Smoking129 (1.8)771 (1.8)0.897Hypertension1,272 (18.2)4,346 (10.2)<0.001Tuberculosis8 (0.1)29 (0.1)0.283Socioeconomic status<0.001 High904 (12.9)4,910 (11.4) Low2,141 (30.6)14,271 (33.0) Middle3,955 (56.5)24,019 (55.6)Healthcare affiliation<0.001 Contributive6,446 (92.1)35,587 (82.4) Not affiliated/Unknown247 (3.5)5,736 (13.3) Subsidized307 (4.4)1877 (4.3)Vaccination status<0.001 Vaccinated4,098 (58.5)16,609 (38.4) Unvaccinated2,902 (41.5)26,591 (61.6) Open table in a new tab Among primary series with non-mRNA vaccine schedules, a primary series with CoronaVac boosted with ChAdOx1-S was the only schedule lacking a mRNA booster that had significant effectiveness to reduce risk of Omicron symptomatic cases. All other schedules with non-mRNA vaccine primary series that yielded statistically significant risk reductions included mRNA vaccines as boosters (Panel B of the Fig. 1). The effectiveness of these booster vaccination schedules against Omicron by age and comorbidities is shown in Panels C-J of the Fig. 1. Previous epidemiological literature has compared the effectiveness of some of the specific schedules we report, and concurred with our findings. Similar to our significant results in the Panel B of the Fig. 1, a study in the Nordic countries found the effectiveness of the primary series of ChAdOx1-S with an mRNA vaccine booster was 27% (95% CI, 3.7 to 50.6), and significantly better than the homologous ChAdOx1-S booster to prevent hospitalizations during Omicron3Andersson N.W. Thiesson E.M. Baum U. et al.Comparative effectiveness of heterologous third dose vaccine schedules against severe covid-19 during omicron predominance in Nordic countries: population based cohort analyses.BMJ. 2023; 382e074325Google Scholar. Another study from the United States reported that the effectiveness 2-4 months since the last dose of the JNJ-78436735 primary series with homologous booster was 29.2% (95% CI, 23.1 to 34.8), and 54.3% (95% CI, 52.2 to 56.3) with the BNT162b2 booster4Accorsi E.K. Britton A. Shang N. et al.Effectiveness of Homologous and Heterologous Covid-19 Boosters against Omicron.N Engl J Med. 2022; 386: 2433-2435Crossref PubMed Scopus (32) Google Scholar. Our results show the JNJ-78436735 homologous booster did not reduce risk of symptomatic Omicron, while a booster with the mRNA-1273 vaccine reduced risk by 43% (see Fig. 1). Our results are similar to a previous Brazilian study5Cerqueira-Silva T. Katikireddi S.V. de Araujo Oliveira V. et al.Vaccine effectiveness of heterologous CoronaVac plus BNT162b2 in Brazil.Nat Med. 2022; 28: 838-843Crossref PubMed Scopus (72) Google Scholar, showing that a primary CoronaVac schedule plus a BNT162b2 booster prevents symptomatic Omicron by 37%, and it increased to 59% with a mRNA-1273 booster. Cerqueira-Silva and col.5Cerqueira-Silva T. Katikireddi S.V. de Araujo Oliveira V. et al.Vaccine effectiveness of heterologous CoronaVac plus BNT162b2 in Brazil.Nat Med. 2022; 28: 838-843Crossref PubMed Scopus (72) Google Scholar in Brazil also found that increasing age decreased the effectiveness of the primary series of CoronaVac with the BNT162b2 booster, for both infection and severity, while Ranzani and col.6Ranzani O.T. Hitchings M.D.T. de Melo R.L. et al.Effectiveness of an inactivated Covid-19 vaccine with homologous and heterologous boosters against Omicron in Brazil.Nat Commun. 2022; 13https://doi.org/10.1038/S41467-022-33169-0Crossref Google Scholar in Brazil found increased effectiveness against symptomatic Omicron with increased age of this schedule, and decreased effectiveness against severity. According to our knowledge, there are no other studies showing the difference in vaccine effectiveness according to age or comorbidity status of other homologous and heterologous mRNA schedules against Omicron. Our study has several limitations. The first limitation is that these data were collected as part of mandatory surveillance efforts directed by national and local health authorities countrywide. Given the size of the pandemic, some covariates may be missing or incomplete, and we had to discard part of the sample for missing data on tests, and excluded patients with incomplete data on vaccination status that would otherwise be included in the analyses. Second, the sample size, while large for heterologous schedules, was relatively smaller compared to other studies with booster schedules. Third, the sensitivity of real-time polymerase chain reaction or rapid antigen tests reduces over time7Dinnes J. Deeks J.J. Berhane S. et al.Rapid, point-of-care antigen and molecular-based tests for diagnosis of SARS-CoV-2 infection.Cochrane Database of Systematic Reviews. 2021; 2021https://doi.org/10.1002/14651858.CD013705.PUB2/INFORMATION/ENCrossref Google Scholar. For this reason, we excluded cases with more than ten days between the test and the start of symptoms. Fourth, genotyping was not available for individuals study samples, and we therefore relied upon on surveillance of variants based on a sample of cases undertaken by the Colombian National Institute of Health (SFigure 1)8Instituto Nacional de Salud . Noticias coronavirus-genoma. 2022 . 〈https://www.ins.gov.co/Noticias/Paginas/coronavirus-genoma.aspx〉 (accessed May 16, 2022).Google Scholar. Our study also has important strengths. First, we were able to assess the effectiveness of a vaccine program applied to the entire resident population from one large city (Bogotá, Colombia). The data thus encompasses all healthcare attentions within Bogotá, where the reporting of COVID-19 tests and vaccination status is a national mandate. Our test-negative design decreases the likelihood of health-seeking bias affecting our estimates, as may be a risk with cohort studies. In conclusion, our study contributes further evidence that homologous mRNA and heterologous boosters have a better effectiveness profile than non-mRNA homologous boosters for the prevention of symptomatic Omicron. Our results concur with previous immunological and epidemiological evidence and situate heterologous and mRNA boosters at the forefront of reducing the burden of disease resulting from COVID-19 and the Omicron variant and subvariants. Thus, we recommend that vaccination with homologous mRNA and heterologous booster schedules be a worldwide priority going forward to prevent Omicron. This paper received funding from Universidad Nacional de Colombia (Bogotá, Colombia), and ALZAK Foundation (Cartagena, Colombia).