A multi-country analysis of COVID-19 hospitalizations by vaccination status

Abstract

Background: Individuals vaccinated against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), when infected, can still develop disease that requires hospitalization. It remains unclear whether these patients differ from hospitalized unvaccinated patients with regard to presentation, coexisting comorbidities, and outcomes.

Methods: Here, we use data from an international consortium to study this question and assess whether differences between these groups are context specific. Data from 83,163 hospitalized COVID-19 patients (34,843 vaccinated, 48,320 unvaccinated) from 38 countries were analyzed.

Findings: While typical symptoms were more often reported in unvaccinated patients, comorbidities, including some associated with worse prognosis in previous studies, were more common in vaccinated patients. Considerable between-country variation in both in-hospital fatality risk and vaccinated-versus-unvaccinated difference in this outcome was observed.

Conclusions: These findings will inform allocation of healthcare resources in future surges as well as design of longer-term international studies to characterize changes in clinical profile of hospitalized COVID-19 patients related to vaccination history.

Funding: This work was made possible by the UK Foreign, Commonwealth and Development Office and Wellcome (215091/Z/18/Z, 222410/Z/21/Z, 225288/Z/22/Z, and 220757/Z/20/Z); the Bill & Melinda Gates Foundation (OPP1209135); and the philanthropic support of the donors to the University of Oxford's COVID-19 Research Response Fund (0009109). Additional funders are listed in the "acknowledgments" section.

Keywords: COVID-19; Translation to population health; comorbidity; descriptive epidemiology; heterogeneity; vaccination.

Graphical abstract

Figure 1

Selection of the analytic sample A database from the Our World in Data initiative with information on country-level vaccination coverage was used in the final step of selection. Pregnant women were excluded from the analysis because previous studies reported that delivery is a common reason for hospital admission in patients with incidental SARS-CoV-2 infections. See also Figure S1.

Figure 2

Manufacturers of vaccines administered to study participants The y axes present the distributions of vaccine manufacturers, as proportions, by country (x axes; the same ordering applies to the three panels). For some participants, information on vaccine manufacturer was available for the second or third doses but not for the first dose (N = 25 and 15, respectively) or for the third dose but not the second dose (N = 18). Only countries with at least 20 participants for whom manufacturer information was available are included in this figure; this criterion was also used for each dose-specific panel. Information on vaccine manufacturer for the fourth dose was available for 22 patients and is not presented here. See also Figure S2 and Table S2.

Figure 3

Country-specific frequencies of the 10 most common symptoms Red bars correspond to data from unvaccinated patients and light blue bars to data from vaccinated patients. The ordering of symptoms (x axes) is the same in all panels; y axes present frequencies as proportions. The 95% confidence intervals are also shown. Only countries with at least 100 patients with data on one or more symptoms are included; the ISO3 code of each country included is presented as the corresponding panel title: CAN, Canada; IND, India; KWT, Kuwait; LAO, Laos; NPL, Nepal; NLD, the Netherlands; NOR, Norway; PAK, Pakistan; ROU, Romania; ESP, Spain; GBR, United Kingdom; USA, United States. See Table S4.

Figure 4

Differences in frequencies of comorbidities between vaccinated and unvaccinated groups Coordinates in the y axis correspond to comorbidities, ordered based on mean frequencies in all countries, and in the x axis, countries, represented by ISO3 codes, are ordered alphabetically. The four different sizes of the squares in the figure relate to the corresponding frequencies in the unvaccinated group (see top of the graph). Colors represent country- and comorbidity-specific numerical differences in frequencies between unvaccinated and vaccinated patients; red tones indicate that a comorbidity was more frequent in unvaccinated patients. Only countries with 100 or more patients with data on at least one comorbidity are presented. Stars indicate when the number of vaccinated or unvaccinated individuals was below 20. A different version of this figure is shown in the Supplementary Appendix (Figure S3) that accounts for frequencies of comorbidities in the unvaccinated group not only in the size of the squares but also in the color; i.e., the other version of this figure presents differences relative to the frequencies in the unvaccinated group rather than absolute differences.

Figure 5

Fatality risks in vaccinated (y axis) and unvaccinated (x axis) patients by country, represented by different colors, and age Patients were grouped in two broad age categories for this figure: those aged between 18 and 60 years, represented by triangles, and patients older than 60 years, represented by circles. The 95% confidence intervals for both the vaccinated (vertical lines) and unvaccinated (horizontal lines) groups are shown. Figure S4 presents a version of this figure where the two axes range from 0 to 0.30, allowing better visualization of data from countries with lower fatality risks, including South Africa and the United Kingdom. See also Tables S7 and S8.