Tuberculosis following two-dose SARS-CoV-2 vaccination with messenger RNA vaccine (BNT162b2) and inactivated virus vaccine (CoronaVac)

Although the waning of antibodies is anticipated after two-dose SARS-CoV-2 vaccination, the cellular response, especially the Th1 cell response that promotes T-cell immunity, has been reported recently.1Shi T. Dai M.X. Liu F.W. et al.Dynamics of immune responses to inactivated COVID-19 vaccination over 8 months in China.J Infect. 2022; Abstract Full Text Full Text PDF Scopus (1) Google Scholar Development of the T-cell and cellular immune response triggers long-term memory with potential cross-pathogen protection – known as trained immunity.2Netea M.G. Joosten L.A. Latz E. et al.Trained immunity: a program of innate immune memory in health and disease.Science. 2016; 352: aaf1098Crossref PubMed Scopus (1393) Google Scholar,3Netea M.G. Dominguez-Andres J. Barreiro L.B. et al.Defining trained immunity and its role in health and disease.Nat Rev Immunol. 2020; 20: 375-388Crossref PubMed Scopus (877) Google Scholar Animal and epidemiological studies4Covian C. Retamal-Diaz A. Bueno S.M. Kalergis A.M. Could BCG vaccination induce protective trained immunity for SARS-CoV-2?.Front Immunol. 2020; 11: 970Crossref PubMed Scopus (62) Google Scholar, 5Escobar L.E. Molina-Cruz A. Barillas-Mury C. BCG vaccine protection from severe coronavirus disease 2019 (COVID-19).Proc Natl Acad Sci USA. 2020; 117: 17720-17726Crossref PubMed Scopus (264) Google Scholar, 6Zhang B.Z. Shuai H. Gong H.R. et al.Bacillus calmette-guerin-induced trained immunity protects against SARS-CoV-2 challenge in K18-hACE2 mice.JCI Insight. 2022; Crossref Scopus (9) Google Scholar showed a cross-protection effect from Bacille Calmette-Guerin (BCG) vaccination on COVID-19 by inducing trained immunity. We hypothesize that SARS-CoV-2 vaccination could also trigger trained immunity and offer protection against tuberculosis (TB) through a similar mechanism. In this population-based real-world outcome study in Hong Kong, we linked territory-wide electronic health records (EHRs) with SARS-CoV-2 vaccination records and applied two epidemiological study designs, case-control study and retrospective cohort study, to investigate the effect of two-dose SARS-CoV-2 vaccination on the occurrence of TB. Matching between EHR and vaccination records was based on anonymized personal identification document numbers. The record-linked EHR database has been used for several population-based pharmacovigilance studies for the SARS-CoV-2 vaccine with proven population representativeness and data accuracy.7Li X. Tong X. Yeung W.W.Y. et al.Two-dose COVID-19 vaccination and possible arthritis flare among patients with rheumatoid arthritis in Hong Kong.Ann Rheum Dis. 2022; 81: 564-568Crossref PubMed Scopus (56) Google Scholar,8Lai F.T.T. Li X. Peng K. et al.Carditis after COVID-19 vaccination with a messenger RNA vaccine and an inactivated virus vaccine: a case-control study.Ann Intern Med. 2022; 175: 362-370Crossref PubMed Scopus (57) Google Scholar We identified the interest of outcome as newly diagnosed TB from the inpatient setting between February 23, 2021, and January 31, 2022, using ICD-9-CM diagnostic codes (010–018). To ensure the TB cases were incident events during the period, patients with a recorded TB diagnosis or TB-related antibiotics prescription (isoniazid or rifampin) were excluded. Cases were further verified by prescription records of isoniazid, rifampin, pyrazinamide and ethambutol or streptomycin within 14 days after hospital admission. In the cohort study, we included all the patients’ records in the linked database and categorized the cohort into two-dose vaccinated or unvaccinated group according to the vaccination status by September 30, 2021. We matched vaccine recipients with unvaccinated individuals by age and sex using maximum ratio matching and followed them up until the occurrence of outcome, death or study end date. Patients with metastatic cancer, age<18 years, with clinical history of TB or TB-related treatment, or with single-dose or heterologous vaccines were excluded. Multi-group Inverse Probability of Treatment Weighting (IPTW) was adopted to ascertain the balance of patient characteristics across groups. Cox Proportional-Hazards model was applied to estimate the hazard ratio (HR). In the nested case-control study, TB cases were 1:10 matched with controls admitted to hospital during the same period but without a diagnosis of TB, using the incidence density sampling with replacement by age, sex, and hospital admission date (±1 day). Multivariable conditional logistic regression was applied to evaluate odds ratio (OR). HR and OR were estimated separately for BNT162b2 and CoronaVac. Subgroup analysis (by age, sex) and a series of sensitivity analyses were conducted. The detailed study design and statistical analysis are shown in Supplementary Methods. The study cohort included 1662,879 unvaccinated individuals, 1320,654 two-dose BNT162b2 vaccine recipients, and 944,331 two-dose CoronaVac vaccine recipients (Supplementary Fig. 1). After IPTW with 1% extreme values trimmed, we obtained a well-balanced cohort with all standard mean difference (SMD) <0.1 except for age (Supplementary Table 1), which was adjusted by Cox regression. During a median follow-up of 178–199 days, incidence of TB in the BNT162b2 group [(1.35 (95% CI: 1.1–1.63) per 10,000-person year] and the CoronaVac group [1.53 (95% CI: 1.23–1.89) per 10,000-person year] were lower than the unvaccinated group [3.47 (95% CI: 3.09–3.88) per 10,000-person year] (Supplementary Table 2). Cox regression showed the adjusted HR was 0.42 (95% CI: 0.31–0.57) for BNT162b2 and 0.51 (95% CI: 0.39- 0.69) for CoronaVac when compared to the unvaccinated group. Age- and sex-stratified Cox regression showed similar associations for both vaccines (Fig. 1). Sensitivity analyses using the recorded diagnosis of TB regardless of TB-related prescription as the outcome definition, considering 30 days washout period for TB occurrence, Fine-Gray regression considering death as a competing risk for TB, or using appendicitis as the negative outcome control, all showed similar findings with the main analysis (Table 1). The case-control study (Supplementary Fig. 2 for cases and controls selection and Supplementary Table 3 for baseline demographics) yielded similar, but a more conservative risk estimate [adjusted OR 0.76 (95% CI: 0.57–1.01) for BNT162b2; 0.74 (95% CI: 0.56–0.99) for CoronaVac] (Fig. 1).Table 1Sensitivity analyses for cohort study.Events (N)Cohorts (N)Time-to-event [days, median (IQR)]Follow-up time (person-years)Incidence (10,000 person-years, 95% CI)Adjusted HR (95% CI)P-valueHospitalized TB regardless of TB-related prescription within 14 daysNone5541,662,879178 (144, 226)849,4966.52(5.99, 7.08)RefBNT162b21311,320,654188 (158, 230)712,629.81.84(1.54, 2.17)0.36 (0.27, 0.47)<0.001CoronaVac130944,331199 (164, 255)541,1182.4(2.01, 2.84)0.45 (0.36, 0.57)<0.00130-day wash out period for TB definitionNone2501,662,879178 (144, 226)849,587.42.94(2.59, 3.32)RefBNT162b2881,320,654188 (158, 230)712,641.81.23(0.99, 1.51)0.44 (0.32, 0.6)<0.001CoronaVac76944,331199 (164, 255)541,134.31.4(1.11, 1.74)0.54 (0.4, 0.74)<0.001Fine-Gray competing risk of death analysisNone2951,662,879178 (144, 226)849,565.43.47(3.09, 3.88)RefBNT162b2961,320,654188 (158, 230)712,638.11.35(1.1, 1.63)0.49 (0.38, 0.62)<0.001CoronaVac83944,331199 (164, 255)541,130.41.53(1.23, 1.89)0.51 (0.4, 0.66)<0.001Negative outcome control (appendicitis)None4681,662,879178 (144, 226)849,519.75.51(5.02, 6.02)RefBNT162b24951,320,654188 (158, 230)712,514.66.95(6.35, 7.58)1.10 (0.94, 1.29)0.222CoronaVac328944,331199 (164, 255)541,041.66.06(5.43, 6.74)1.01 (0.87, 1.18)0.859 Open table in a new tab Hong Kong is among the few jurisdictions that implemented two types of SARS-CoV-2 vaccines with established territory-wide vaccine safety surveillance. From both mRNA and inactivated virus vaccine technology platforms, we observed a significantly lower risk of incident TB among people who received two-dose vaccines. The overall estimated relative risk reduction was 49–58% in the cohort analysis and 24–26% in the case-control analysis. Consistent findings from sensitivity analyses further supports the trained immunity theory, and it is likely that the cross-pathogen protection could be sustained for at least 6 months, according to the median follow-up period of the cohort study. Long-term boosting of innate immune responses by live vaccines, such as BCG, could potentially induce heterologous protection against infections through epigenetic, transcriptional, and functional reprogramming of innate immune cells.9Netea M.G. Giamarellos-Bourboulis E.J. Dominguez-Andres J. et al.Trained immunity: a tool for reducing susceptibility to and the severity of SARS-CoV-2 infection.Cell. 2020; 181: 969-977Abstract Full Text Full Text PDF PubMed Scopus (288) Google Scholar Therefore, it was proposed that the induction of trained immunity might represent an important tool for reducing susceptibility to and severity of SARS-CoV-2,9Netea M.G. Giamarellos-Bourboulis E.J. Dominguez-Andres J. et al.Trained immunity: a tool for reducing susceptibility to and the severity of SARS-CoV-2 infection.Cell. 2020; 181: 969-977Abstract Full Text Full Text PDF PubMed Scopus (288) Google Scholar which was recently proved in an animal study with SARS-CoV-2 challenge.6Zhang B.Z. Shuai H. Gong H.R. et al.Bacillus calmette-guerin-induced trained immunity protects against SARS-CoV-2 challenge in K18-hACE2 mice.JCI Insight. 2022; Crossref Scopus (9) Google Scholar Our results, consistent with the trained immunity theory, warrants further pathogenesis and epigenetic investigation. Notably, our observation relating to the cross-pathogen protection is not specific to mRNA or inactivated virus vaccine platform. This indicates that the trained immunity might involve several cell-pathogenesis cross-talks and regulations, and the potential of whole-microorganism vaccines as an important tool for reducing the susceptibility of SARS-CoV-2. Despite several limitations inherently associated with EHR-based real-world outcome studies, our study has significant public health implications, particularly for low-and low-middle-income economies with dual threats from high prevalent TB and uncontrolled COVID-19 due to low uptake of SARS-CoV-2 vaccines.10Duan Y. Shi J. Wang Z. Zhou S. Jin Y. Zheng Z.J. Disparities in COVID-19 Vaccination among low-, middle-, and high-income countries: the mediating role of vaccination policy.Vaccines. 2021; 9 (Basel): 905Crossref PubMed Scopus (30) Google Scholar Potential additional benefits of SARS-CoV-2 vaccination should be made known to the public to overcome vaccine hesitancy; and to policymakers, to facilitate feasible and cost-effective vaccination programs for COVID-19 and TB control. This study was approved by the Institutional Review Board of the University of Hong Kong / Hospital Authority Hong Kong West (UW 21–149 and UW 21–138) and the Department of Health Ethics Committee (LM 21/2021). Data are not available as the data custodians (the Hospital Authority and the Department of Health of Hong Kong SAR) have not given permission for sharing due to patient confidentiality and privacy concerns. Local academic institutions, government departments, or non-governmental organizations may apply for access to data through the Hospital Authority's data sharing portal (https://www3.ha.org.hk/data). Professor Wong had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Concept and design: XL, KP, ICKW Drafting of the manuscript: XL, KP Data acquisition and management: CSLC, FTTL, EYFW, XL, CKW, EWC, ICKW Statistical analysis: KP, FC, XL Clinical investigators: DCLL, MSMI, CSL Interpretation of data: all authors Critical revision of the manuscript for important intellectual content: all authors Administrative, technical, or material support: ICKW, EWC Supervision: XL, ICKW Food and Health Bureau, the Government of the Hong Kong Special Administrative Region (Ref: COVID19F01).