SARS-CoV-2 vaccine breakthrough infections in Virginia, January 17, 2021 – June 30, 2021

Abstract

As COVID-19 vaccines moved from the controlled environment of clinical trials to use in real-world settings, it has been important to evaluate vaccine effectiveness. A retrospective cohort study was designed to identify cases of SARS-CoV-2 infection that occurred between January 17-June 30, 2021 in fully vaccinated Virginia residents. Of the fully vaccinated population of Virginia at the end of the study period (N = 4,271,505), 2445 (0.057 %) were reported to have experienced a vaccine breakthrough infection. Of those, 183 (7.5 %) were reported to have been hospitalized for COVID-19 and 53 (2.2 %) died from COVID-19. There were significant differences in vaccine effectiveness over time between both mRNA vaccines and the Janssen vaccine. Increasing age, pre-existing medical conditions, and male sex were associated with severe outcomes (hospitalization or death). Persons at greater risk for severe outcomes should continue to take precautions to prevent SARS-CoV-2 infection, even if fully vaccinated.

1. Introduction

COVID-19 vaccines that have been available for use in the United States have demonstrated vaccine efficacies ranging between 66.1 and 95.0 % in initial clinical trials [1], [2], [3]. As vaccines moved from the controlled environment of a clinical trial to use in real-world settings, it has been important to evaluate vaccine effectiveness. No vaccine is expected to be 100 % effective at preventing illness and vaccine breakthrough infections are expected, though vaccines should continue to provide a high level of protection against serious clinical outcomes when used outside of a clinical trial. The Virginia Department of Health (VDH) analyzed SARS-CoV-2 vaccine breakthrough infections reported to occur between January 17, 2021 (the first date a Virginia resident could meet the definition of fully vaccinated1 ) and June 30, 2021 to characterize breakthrough infections in Virginia and examine risk factors and outcomes associated with these infections.

2. Material and methods

A retrospective cohort study was designed to identify cases of COVID-19 that occurred between January 17-June 30, 2021 in Virginia residents where SARS-CoV-2 RNA or antigen was detected in a respiratory specimen collected from a person ≥ 14 days after they have completed all recommended doses of a U.S. Food and Drug Administration (FDA)-authorized or approved COVID-19 vaccine. Persons who had a previous positive COVID-19 nucleic acid amplification test recorded in the Virginia Electronic Disease Surveillance System (VEDSS) within 90 days preceding the positive test that qualified them to meet the definition of a SARS-CoV-2 vaccine breakthrough infection case were excluded from analysis, as the qualifying test could be consistent with a person persistently testing positive after acute infection [4], [5], [6]. Available case information in VEDSS and the Virginia Immunization Information System (VIIS) were analyzed to understand patterns of SARS-CoV-2 vaccine breakthrough infection by age, sex, race and ethnicity, vaccine type and product, the influence of time (used as a proxy for waning immunity), reported pre-existing conditions, and infecting SARS-CoV-2 variant. Follow up time, defined as the time between a person meeting the definition of fully vaccinated and the end of the study period, and the time to vaccine breakthrough, defined as the time between meeting the definition of fully vaccinated and the specimen collection date associated with their vaccine breakthrough infection in VEDSS, were calculated. Frequency distributions were generated to summarize demographic characteristics, infecting variants, pre-existing conditions, and selected health outcomes (e.g., symptomatic illness, hospitalization, and death) within the study population. Chi-square and Fisher Exact tests analyzed differences in observed and expected rates of vaccine breakthrough. Published reports of vaccine efficacy [1], [2], [3] and aggregate data on the fully vaccinated population in Virginia on June 30, 2021 were used to generate expected rates of vaccine breakthrough. The Kaplan-Meier estimator was used to evaluate the effect of time on the risk of experiencing a vaccine breakthrough infection. Multivariable logistic regression analyzed the associations between potential risk factors (age, sex, race and ethnicity, vaccine product, time to vaccine breakthrough, the presence of pre-existing conditions, and variant type) on severe outcomes (hospitalization or death). R (version 4.2.0) was used to conduct all analyses.

3. Results

Of the fully vaccinated population of Virginia at the end of the study period (N = 4,271,505), 2445 (0.057 %) were reported to have experienced a vaccine breakthrough infection. Reported vaccine breakthrough cases ranged in age from 13 to 102 (mean 54.1, SD 19.5), were 60.0 % (1468/2445) female, 61.8 % (1510/2445) Non-Hispanic White, and 56.7 % (1387/2445) received the Pfizer-BioNTech vaccine. A summary of breakthrough cases stratified by age, sex, race/ethnicity, and vaccine product is presented in Table 1 .

Table 1.

SARS-CoV-2 Vaccine Breakthrough Infections, Virginia, January 17, 2021 – June 30, 2021.

	Pfizer-BioNTech			Moderna			Janssen
	Vaccine Breakthrough Cases	Population at Risk	Observed Vaccine Breakthrough Rate (%)	Vaccine Breakthrough Cases	Population at Risk	Observed Vaccine Breakthrough Rate (%)	Vaccine Breakthrough Cases	Population at Risk	Observed Vaccine Breakthrough Rate (%)
Overall	1387	2,333,839	0.059	626	1,621,228	0.039	432	316,438	0.136
Age Group
12–17 years	21	230,084	0.009
18–30 years	164	382,910	0.043	76	236,940	0.032	67	58,746	0.114
31–50 years	433	675,516	0.064	188	453,823	0.041	138	103,922	0.133
51–64 years	293	534,412	0.055	150	410,199	0.037	117	93,803	0.125
65 + years	476	510,917	0.093	212	520,266	0.041	110	59,967	0183
Sex
Male	506	1,051,406	0.048	255	724,089	0.035	205	166,651	0.123
Female	871	1,274,899	0.068	371	892,049	0.042	226	148,589	0.152
Unknown	10	7534	0.133	0	5090	0.000	1	1198	0.083
Racial or Ethnic Group
Asian or Pacific Islander	84	251,787	0.033	29	115,877	0.025	25	20,105	0.124
Black	216	335,153	0.064	94	246,946	0.038	69	39,045	0.177
Latino	37	207,187	0.018	12	109,844	0.011	4	26,206	0.015
Native American	0	11,015	0.000	1	6,496	0.015	1	1,374	0.073
White	822	1,224,723	0.067	423	985,407	0.043	265	191,528	0.139
Other	91	137,994	0.066	26	39,135	0.066	25	10,209	0.245
Unknown	137	165,980	0.082	41	117,523	0.035	43	27,971	0.154

The median follow up time was 61 days (IQR 32–84). Seventy-five percent of reported vaccine breakthrough cases occurred within the first 65 days after becoming fully vaccinated (median 38, IQR 18–65). Fig. 1 shows observed vaccine efficacy over time by vaccine product type. There were significant differences in vaccine effectiveness over time between both mRNA vaccines and the Janssen vaccine (Pfizer-BioNTech:Janssen, P < 0.001, Moderna:Janssen, P < 0.001), but there was no significant difference between the mRNA vaccines (P = 0.5).

Fig. 1.

SARS-CoV-2 Vaccine Breakthrough Infections by Vaccine Product over Time.

Of the 2445 cases, symptomatic illness was reported in 62.7 % (1532); 7.5 % (1 8 3) were known to be hospitalized for COVID-19 and 2.2 % (53) cases died from COVID-19. Symptomatic illness, hospitalization, or death were not reported for the remaining 667 cases. The presence of pre-existing health conditions that increase the risk of experiencing more severe outcomes from COVID-19 were reported by 46.3 % (1132) of cases and current pregnancy was reported by 0.5 % (12) of cases. Whole genome sequencing of a positive SARS-CoV-2 viral sample was performed on 354 cases; of these, 49.4 % (1 7 5) were Alpha, 22.3 % (79) were Delta, and 28.2 % (1 0 0) were other variants.

In the regression model for hospitalizations, increasing age (adjusted odds ratio [OR] 1.08, 95 % CI 1.07–1.10) and the presence of pre-existing medical conditions (OR 5.59, 95 % CI 2.54–12.31) were both significantly associated with being hospitalized for COVID-19 (P < 0.001). In this model, sex, and race and ethnicity were not significantly associated with COVID-19 hospitalization. In a smaller hospitalization model that included only cases where the infecting SARS-CoV-2 variant was known, males were found to be at increased risk of hospitalization (OR 2.84, 95 % CI 1.19–6.77); the infecting variant was not significant. In the deaths model, only increasing age (OR 1.10, 95 % CI 1.07–1.13) was significantly associated (P < 0.001) with dying from COVID-19. For every 10-year increase in age, there was a 2.2 times greater risk of being hospitalized and a 2.6 times greater risk of dying from a SARS-CoV-2 vaccine breakthrough infection. Race and ethnicity, vaccine product, and time to vaccine breakthrough were not significant in any model. The regression models for COVID-19 hospitalizations and deaths are summarized in Table 2 .

Table 2.

Summary of Logistic Regression Models of SARS-CoV-2 Vaccine Breakthrough Hospitalizations and Deaths, Virginia, January 17, 2021 – June 30, 2021.

	Hospitalizations			Hospitalizations Subset with Sequenced Variant Data			Deaths
Sample size	1715			285			1605
Variable	β	Std. Error	P	β	Std. Error	P	β	Std. Error	P
Constant	−9.088	0.644	<0.001	−9.157	1.440	<0.001	−11.340	1.323	<0.001
Age	0.078	0.007	<0.001	0.078	0.016	<0.001	0.094	0.015	<0.001
Sex, Male	0.352	0.193	0.068	1.043	0.444	0.019	0.376	0.349	0.280
Race/Ethnicity, Other than Non-Hispanic White	0.291	0.216	0.177	0.481	0.461	0.297	−0.060	0.422	0.887
Pre-Existing Medical Condition	1.721	0.403	<0.001	1.857	1.062	0.080	1.344	0.744	0.071

4. Discussion

The United States Census Bureau estimated that Virginia had a total population of 8,631,393 in 2020 [7]. By this estimate, around half of all Virginians were fully vaccinated by the end of the study period. Between January 17-June 30, 2021, 140,664 confirmed cases2 of COVID-19 were reported in VEDSS. Overall, rates of reported SARS-CoV-2 vaccine breakthrough infection were very low (<1% of the fully vaccinated population; 1.7 % of confirmed cases). This is similar to other SARS-CoV-2 vaccine breakthrough reports that relied on passive reporting of cases to public health departments during a comparable timeframe [8], [9], [10], [11]. Females were overrepresented in this study population, which is similar to other early U.S. reports on SARS-CoV-2 vaccine breakthrough [8], [9], [10], and might reflect differences in vaccine-seeking and healthcare-seeking behaviors between the sexes during this time. Even though females were overrepresented among fully vaccinated people and vaccine breakthrough cases, they were not at increased risk for severe outcomes. Males were found to be at increased risk for COVID-19 hospitalization in one model, but were not consistently found to be at increased risk for severe outcomes across models. Further study is needed to examine if males are truly at increased risk of being hospitalized due to SARS-CoV-2 vaccine breakthrough infection.

Overall rates of symptomatic illness (62.7 %), hospitalization (7.5 %), and death (2.2 %) in this study are consistent with other reports of disease severity in SARS-CoV-2 vaccine breakthrough cases detected through passive reporting [8], [9], [10], [12]. Increasing age and the presence of certain pre-existing medical conditions are well-recognized risk factors for severe outcomes from COVID-19. While a wide range of ages experienced vaccine breakthrough infections in this study, older people were consistently at increased risk for hospitalization and death. The presence of pre-existing medical conditions was a significant risk factor in the larger COVID-19 hospitalization model, but not the other two models; this is likely due to the smaller sample sizes in these models. Pre-existing conditions in this study were self-reported and analyzed as a dichotomous variable; this likely introduced some degree of self-reporting bias and did not permit further analysis of pre-existing condition severity and their influence on health outcomes.

All three COVID-19 vaccines that were available during the study period (i.e., Pfizer-BioNTech, Moderna, Janssen) were effective at preventing severe outcomes, however, consistent with other reports [13], [14] the effectiveness of the Janssen vaccine product at protecting against reported infection declined more rapidly than the mRNA vaccine products.

This study has several limitations. Firstly, persons with mild or asymptomatic infections may be less likely to seek testing. Additionally, at-home COVID-19 tests have become increasingly available and these results are not always reported to public health. This means that reported breakthrough infections are likely an underrepresentation of all vaccine breakthrough infections and explains some of the difference between the rates of COVID-19 vaccine breakthrough reported in initial clinical trials (5.0–33.9 %) and in this study (<1%) during a similar 60-day follow-up period. Secondly, detailed case information is typically collected by public health during a case interview. Persons who were not interviewed in detail due to staffing shortages during COVID-19 surges and persons who did not consent to be interviewed likely have incomplete data. Further, some people may have been hospitalized subsequent to an interview; as hospitalization information is mostly collected during the interview process, the number of hospitalized patients might be greater than reported here. Hospitalization data might be more accurate in the variant-typed subgroup, as those persons were more likely to be reinterviewed if sequencing data showed an emerging variant. This might be why a significant association with males being at increased risk for COVID-19 hospitalization compared to females was detected in the smaller hospitalization model that only included cases with genomic sequencing, but not in the larger hospitalization model. Thirdly, as with any observational study, there are additional variables that could have influenced health outcomes that were either not available for analysis or could not be measured as precisely as desired; as such, residual confounding might be present. Examples of these additional variables include, but are not limited to, a person’s socioeconomic status, healthcare-seeking behavior, vaccination-seeking behavior, and coinfection with other respiratory pathogens.

This report provides valuable insights as to which populations were impacted by SARS-CoV-2 vaccine breakthrough infections during early 2021, the severity of reported illness these people experienced, and identified risk factors associated with severe health outcomes. While vaccine breakthrough infections occurred at low rates, hospitalizations and deaths did occur at non-negligible rates. To identify changes in populations bearing the greatest burden of disease and risk factors associated with severe outcomes, continued monitoring of SARS-CoV-2 vaccine breakthrough case trends is necessary. This type of analysis can also complement vaccine effectiveness studies by estimating the protective effect of COVID-19 vaccination at a population-level.

5. Conclusions

While reported SARS-CoV-2 vaccine breakthrough infections occurred rarely during this study period, they can have serious outcomes. Continued monitoring of COVID-19 vaccine breakthrough infections is important to monitor trends over time and can complement vaccine effectiveness studies. Persons at greater risk for severe outcomes should continue to take precautions to prevent SARS-CoV-2 infection, even if fully vaccinated.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The authors do not have permission to share data.