Hybrid Immunity Provides Protective Advantage Over Vaccination or Prior Remote Coronavirus Disease 2019 Alone

Abstract

Background

The protective efficacy of prior coronavirus disease 2019 (COVID-19) with or without vaccination remains unknown. This study sought to understand if 2 or more messenger RNA (mRNA) vaccine doses provide additional protection in patients with prior infection, or if infection alone provides comparable protection.

Methods

We conducted a retrospective cohort study of the risk of COVID-19 from 16 December 2020 through 15 March 2022, among vaccinated and unvaccinated patients of all ages with and without prior infection. A Simon-Makuch hazard plot illustrated the incidence of COVID-19 between groups. Multivariable Cox proportional hazards regression was used to examine the association of demographics, prior infection, and vaccination status with new infection.

Results

Among 101 941 individuals with at least 1 COVID-19 polymerase chain reaction test prior to 15 March 2022, 72 361 (71.0%) received mRNA vaccination and 5957 (5.8%) were previously infected. The cumulative incidence of COVID-19 was substantially higher throughout the study period for those previously uninfected and unvaccinated, and lowest for those previously infected and vaccinated. After accounting for age, sex, and the interaction between vaccination and prior infection, a reduction in reinfection risk was noted during the Omicron and pre-Omicron phases of 26% (95% confidence interval [CI], 8%–41%; P = .0065) to 36% (95% CI, 10%–54%; P = .0108), respectively, among previously infected and vaccinated individuals, compared to previously infected subjects without vaccination.

Conclusions

Vaccination was associated with lower risk of COVID-19, including in those with prior infection. Vaccination should be encouraged for all including those with prior infection, especially as new variants emerge and variant-specific booster vaccines become available.

COVID-19 mRNA vaccination was associated with lower risk of COVID-19, including in those with prior COVID-19. Vaccination should be encouraged for those with remote prior infection, especially as new variants emerge and variant-specific booster vaccines become available.

Graphical Abstract

Graphical Abstract.

This graphical abstract is also available at Tidbit: https://tidbitapp.io/tidbits/COVID-19-vaccination-is-boosts-more-protective-immunity-in-previously-infected/update

The coronavirus disease 2019 (COVID-19) pandemic, caused by the severe acute respiratory syndrome coronavirus 2 (SARS–CoV-2), continues worldwide with hundreds of millions of infections and millions of deaths. Since December 2020, 4 COVID-19 vaccines received Emergency Use Authorization in the United States (US), with 2 fully licensed by the US Food and Drug Administration. These vaccines were developed on previously studied novel platforms followed by successful completion of phase 1, 2, and 3 clinical trials [1–6]. The vaccines were shown to provide >90% protective efficacy against the prevalent strain (predominantly the ancestral strain) circulating during the clinical trials. During the ensuing 18 months, >257 million Americans received at least 1 dose of these vaccines, and >70% of the population over 5 years of age are considered fully vaccinated [7]. However, during the same time, there has also been further evolution of SARS-CoV-2 with the emergence of new variants of concern (VOCs), specifically Delta, Omicron, and subvariants of Omicron [7]. Concurrent with the emergence of these variants, real-world data showed that protection from these vaccines declined [8–10], particularly for mild disease [11–14]. Due to the confluence of unvaccinated individuals, waning immunity, and new variants, a large proportion of the US population has had COVID-19. A seroprevalence study by the Centers for Disease Control and Prevention showed that as of February 2022, approximately 75% of children and adolescents had serologic evidence of prior infection, while the overall seroprevalence rate in the US population was 57.7% (95% confidence interval [CI], 57.1%–58.3%]) [15]. Pediatric seroprevalence was estimated to be 86.3% (95% CI, 85.7%–86.8%) by August 2022 [16]. Prior studies have also shown that immunity afforded by SARS-CoV-2 infection is similar to vaccine-mediated immunity and lasts for a similar duration [17]. Additionally, hybrid immunity from vaccination in previously infected individuals provides better protection from hospitalization than immunity from infection alone [18, 19]. Recent studies show that compared to unvaccinated persons, vaccinated individuals have a very robust immune response to subsequent infection with broader cross-reactive antibodies against other VOCs [20]. Similarly, neutralizing antibody responses to BA.4 or BA.5 are higher in previously vaccinated individuals with BA.1 infection than in unvaccinated individuals with previous BA.1 infection [21]. However, the protective efficacy of prior infection against reinfection overall and the amount of additional protection vaccination provides in previously infected individuals remains unclear, especially as new VOCs emerge. Therefore, COVID-19 vaccination is recommended to all including previously infected persons.

The goal of this study is to understand if COVID-19 vaccination with 2 or more doses of messenger RNA (mRNA) vaccine provides additional protection in patients with prior COVID-19, or if infection alone provides comparable protection against reinfection. We compared the incidence of COVID-19 in a large regional cohort of vaccinated patients with and without a history of prior infection. We also compared the incidence of COVID-19 in previously infected patients with and without subsequent vaccination. Our hypothesis was that 2 or more mRNA COVID-19 vaccine doses provide additional protection in individuals with a prior history of COVID-19. This will hopefully add to the knowledge of whether COVID-19 vaccine series or boosters should be recommended in patients with prior COVID-19.

METHODS

Subjects and Setting

This is a retrospective study of individuals who underwent polymerase chain reaction (PCR) testing for suspected SARS-CoV-2 infection at the Mayo Clinic and Mayo Clinic Health System. This study was reviewed and approved by the Mayo Clinic Institutional Review Board (IRB 21–000967) as a minimal risk study. Patients were excluded if they lacked research authorization per state statutes. The study population consisted of 101 941 individuals of any age empaneled to a primary care provider at any Minnesota or Wisconsin Mayo Clinic site with a record of at least 1 COVID-19 PCR test prior to the end of the study period. COVID-19 vaccine was first available on 16 December 2020, and rolled out in a staged fashion previously described [22]. Vaccination and COVID-19 test positivity were documented in the electronic health record via direct entry or import from state immunization information systems as previously described [23]. Patients were considered fully vaccinated if they had received 2 or more COVID-19 mRNA vaccine doses of either the BNT162b2 (Pfizer-BioNTech) or mRNA-1273 (Moderna) vaccine. Very few patients received heterologous vaccination and were categorized as fully vaccinated based upon evidence of equivalent or slightly improved immunologic response with heterologous boosting among mRNA vaccinations [24]. Prior evidence shows comparative vaccine effectiveness of the 2 brands of mRNA vaccine [25]. Patients were censored on the date of the first dose for those who received Janssen COVID-19 vaccine, for whom vaccine manufacturer information was missing, or who received a vaccine other than Moderna or Pfizer. Subjects who only had a single dose of either BNT162b2 or mRNA-1273 vaccine were included as unvaccinated because only 216 of 5957 (3.6%) had a single mRNA vaccination and, of those, only 13 had prior COVID-19 and a reinfection. These were felt to be too small to be assessed as fully vaccinated or evaluated separately. We evaluated the vaccine effectiveness in these groups by measuring the number of subjects with a positive COVID-19 PCR test during the study period, from 16 December 2020 until 15 March 2022. The latter date was chosen due to the widespread availability of home antigen tests after this date and inability to accurately capture positive cases. Subjects were included in the previously infected group if they had a positive COVID-19 PCR test prior to 16 December 2020. Any positive test within 90 days of that prior positive test was censored to allow for a subsequent infection to be unrelated. Individuals with a newly positive COVID-19 PCR between their first mRNA vaccine and day 14 after their second dose were excluded from analysis. Janssen COVID-19 vaccine recipients were excluded due to the low number of individuals in this group, limiting power to draw conclusions.

Statistical Analysis

Baseline demographics of the study cohort are presented as frequency and percentage for categorical variables and compared between groups using Pearson χ² test. Continuous variables are presented using mean and standard deviation and compared between groups using 2-sample t test. The primary outcome was time to new COVID-19 during the study duration. Incidence of COVID-19 was compared in 4 cohorts of patients, which included (1) previously uninfected and unvaccinated, (2) previously uninfected and vaccinated, (3) previously infected and unvaccinated, and (4) previously infected and vaccinated, similar to previously published work [26]. To illustrate the incidence of COVID-19 between groups, a Simon-Makuch hazard plot was created [27]. Subjects were censored at the time of death or if their empaneled status changed at our institution.

Multivariable Cox proportional hazards regression was used to examine the association of baseline demographics with new COVID-19. These results are presented as hazard ratios (HRs) and corresponding 95% CIs. For those patients who changed vaccination status during the study duration, a time-dependent covariate was included within the Cox model. To study duration effects for pre- and post-Omicron, an interaction term was also included in the Cox model that allowed estimation of the effect of prior COVID-19 and vaccination separately for each time period. As prior work had shown a strong interaction between prior COVID-19 and vaccination [26], this interaction was also included in multivariable models. To allow for ease of interpretation, HRs were derived based on estimates from the model, which included baseline covariates, time-dependent effects, and interactions. These HRs are presented alongside the results from model covariates. SAS version 9.4 software was used for analyses and 2-sided P values <.05 were considered statistically significant.

RESULTS

Of the 101 941 subjects in the study, 56 116 (55.0%) were female, 5957 (5.8%) had prior COVID-19, and 72 361 (71.0%) received either the BNT162b2 (Pfizer-BioNTech) or mRNA-1273 (Moderna) vaccine by the end of the study period. Vaccinated subjects were more likely to be older (P < .0001) and female (P < .0001) (Table 1). Among subjects with prior COVID-19, the median interval between infection and the beginning of the study period was 40 days, with an interquartile range of 24–88 days. There were no significant demographic differences between subjects with and without prior infection. Test positivity rates at Mayo Clinic sites in the Midwest reflected expected increases during the surges in the pandemic (Figure 1).

Table 1.

Study Subject Characteristics by Infection and Vaccination Status

Characteristics	Previously Infected (n = 5957)	Not Previously Infected (n = 95 984)	P Value	Vaccinateda (n = 72 361)	Not Vaccinated (n = 29 580)	P Value
Age, y, mean ± SD	38.7 ± 19.7	39.5 ± 24.0	.23	44.2 ± 22.5	27.7 ± 22.7	<.0001
Sex, No. (%)			.70			<.0001
Female	3242 (54.4)	52 874 (55.1)		41 622 (57.5)	14 494 (49.0)
Male	2715 (45.6)	43 103 (44.9)		30 733 (42.5)	15 085 (51.0)
Unknown	0 (0.0)	7 (0.0)		6 (0.0)	1 (0.0)

Abbreviation: SD, standard deviation.

^aFully vaccinated was defined as ≥2 doses of either the BNT162b2 (Pfizer-BioNTech) or mRNA-1273 (Moderna) vaccine.

Figure 1.

Test positivity percentage by date (everyone in the Midwest).

The cumulative incidence of COVID-19 during the study period is reflected in Figure 2. We observed the highest cumulative incidence in those without prior infection or vaccination, and the lowest incidence in subjects with both prior infection and vaccination. Overall, COVID-19 rates were very low in all groups except for the group without prior infection or vaccination prior to the Omicron wave. From the late Delta phase and the first several months of the Omicron phase, subjects with prior infection who remained unvaccinated experienced lower cumulative infection rates than vaccinated individuals without prior infection, but higher infection rates than previously infected subjects who were vaccinated. Despite an increase in infection incidence for all groups during the Omicron phase, the difference in cumulative incidence persisted between groups.

Figure 2.

Cumulative incidence of coronavirus disease 2019 (COVID-19) among the 4 groups: without prior infection or vaccination (orange line); with prior infection and vaccinated (gray line); with prior infection and unvaccinated (blue line); without prior infection and vaccinated (yellow line).

Both vaccination and prior COVID-19 were independently associated with lower rates of infection during the study period. This association was stronger during the pre-Omicron phase and was not modified by age or sex (Table 2).

Table 2.

Factors Associated With Coronavirus Disease 2019 in the Pre- and Post-Omicron Phase

Characteristic	Unadjusted HR (95% CI)	P Value	Adjusted HR (95% CI)	P Value
Prior COVID-19
Pre-Omicron phase	0.17 (.15–.20)	<.0001	0.13 (.10–.16)	<.0001
Omicron phase	0.72 (.65–.80)	<.0001	0.76 (.63–.91)	.0032
Vaccinationa
Pre-Omicron phase	0.36 (.34–.37)	<.0001	0.35 (.34–.36)	<.0001
Omicron phase	0.81 (.77–.85)	<.0001	0.81 (.77–.85)	<.0001
Age (per 10 y)	0.91 (.91–.92)	<.0001	0.97 (.96–.97)	<.0001
Male sex	1.02 (.99–1.05)	.13	0.97 (.95–1.00)	.050
Previously infected: vaccinationa interaction
Pre-Omicron phase	…		1.84 (1.31–2.59)	.0004
Omicron phase	…		0.91 (.73–1.14)	.43
Associations derived from the model after adjusting for interactions
Pre-Omicron phase
No vaccination for those with prior COVID-19	…		0.13 (.10–.16)	<.0001
Vaccination for those without prior COVID-19 vs neither vaccination nor prior COVID-19	…		0.35 (.34–.36)	<.0001
Vaccination for those with prior COVID-19 vs prior COVID-19	…		0.64 (.46–.90)	.011
Vaccination for those without prior COVID-19 vs no vaccination and prior COVID-19	…		2.70 (2.14–3.42)	<.0001
Omicron phase
No vaccination for those with prior COVID-19	…		0.76 (.63–.91)	.0032
Vaccination for those without prior COVID-19 vs neither vaccination nor prior COVID-19	…		0.81 (.77–.85)	<.0001
Vaccination for those with prior COVID-19 vs prior COVID-19	…		0.74 (.59–.92)	.0065
Vaccination for those without prior COVID-19 vs no vaccination and prior COVID-19	…		1.07 (.89–1.28)	.46

Abbreviations: CI, confidence interval; COVID-19, coronavirus disease 2019; HR, hazard ratio.

^aFully vaccinated was defined as ≥2 doses of either the BNT162b2 (Pfizer-BioNTech) or mRNA-1273 (Moderna) vaccine.

We observed a significant interaction between prior COVID-19 and vaccination during the pre-Omicron phase showing a reduced risk of reinfection for vaccinated subjects with prior infection. After accounting for age, sex, and the interaction between vaccination and prior infection, we found a reduction in risk of reinfection during the Omicron and pre-Omicron phases of 26% (95% CI, 8%–41%; P = .0065) to 36% (95% CI, 10%–54%; P = .0108), respectively, among previously infected individuals who were subsequently vaccinated compared to previously infected but unvaccinated subjects (Table 2).

DISCUSSION

We found that COVID-19 vaccination with 2 or more doses of mRNA vaccine (BNT162b2 or mRNA-1273) was associated with a lower risk of subsequent infection for patients with and without prior COVID-19. This supports the recommendation to receive vaccination postinfection.

We found that during the Omicron-predominant period, cumulative infection rates were highest in those without either prior infection or vaccination, and lowest in those with hybrid immunity afforded by prior infection and subsequent vaccination. In the pre-Delta phase, vaccinated but previously uninfected individuals had a lower cumulative incidence of COVID-19 than those with prior COVID-19 only, suggesting that ancestral vaccines were more protective against pre-Delta variants (Figure 2). Differences in infection risk between the 3 groups with prior infection and/or vaccination became apparent during the period when Delta was the predominant circulating VOC (by mid-October 2021, day 205; Figure 1). During this time frame, the previously infected but unvaccinated group demonstrated a lower cumulative risk of infection than the vaccinated group without prior infection, but a higher risk than the hybrid immunity group (Figure 2). From the late Delta through the early Omicron phase, the same observation was noted. The reduction in the cumulative risk of Delta or Omicron infections with 2 or more doses of the monovalent mRNA vaccines in this previously infected cohort is notable as the cohort had COVID-19 prior to 16 December 2020 (pre–Delta VOC). Our study did not specifically evaluate time from prior COVID-19 to vaccination and its impact on protection against reinfection. Rates of reinfection are expected to be lower early after COVID-19 and potentially improve with vaccination when immunity wanes.

During the early Omicron period in January 2022, cumulative infection rates for unvaccinated people with prior COVID-19 approximated the rates for vaccinated people without prior infection, raising the question of whether vaccination provided any benefit for this group. However, as the Omicron period progressed, vaccination in those with prior COVID-19 provided a protective effect compared to those with prior COVID-19 alone (HRs of 0.64 and 0.74 in the pre-Omicron and Omicron periods, respectively). Overall, although the infection rates were low in both groups prior to Omicron, there was a statistically significant reduction in the cumulative risk of infection during both pre-Omicron and Omicron time periods with vaccination.

Several studies have estimated the protective effect of prior infection at approximately 80%–84% for preventing reinfection [17, 28, 29]; however, these studies are all limited by being conducted prior to the emergence of recent variants. Hansen et al found that in a large Danish population cohort, previous infection was <50% effective at protecting individuals aged >65 years from reinfection [28]. These studies were also done prior to the availability of vaccines, so they did not address the role of vaccination in individuals with natural immunity.

Studies have attempted to quantify the additional benefit, if any, of vaccination for those with prior COVID-19. One multistate study in the US of hospitalized patients with COVID-19–like illness found that the odds of laboratory-confirmed COVID-19 were higher (adjusted odds ratio, 5.49 [95% CI, 2.75–10.99]) among previously infected but unvaccinated persons compared to previously uninfected but vaccinated patients during the pre-Delta and Delta periods [30]. This suggests that perhaps vaccination provided better protection than prior infection for COVID-19 hospitalization but is limited by the lack of information on absolute risk of infection or hospitalization for either group. Lack of inclusion of patients with poor access to care and therefore preventive measures such as vaccination could have added additional bias in that study. However, observational studies in large cohorts of previously infected individuals with and without vaccination have generally found that while the absolute risk of reinfection within 6–12 months is quite low, vaccination following infection extends additional significant protection lasting >12 months after the initial infection [12, 17, 26, 31–33]. Vaccination is unlikely to impact recently infected individuals; however, given the lack of a clear correlate of protection and its decline, determination of accurate timing of subsequent vaccination may be challenging.

This study adds to the evolving literature on the impact of vaccination on infection risk among individuals with a prior COVID-19 and is concordant with the findings of others [34–38]. The methodology used was intentionally comparable to the study of healthcare personnel by Shrestha et al [26], in order to provide a similar analysis and outcomes in >101 000 individuals from the general population of all ages to increase the relevance for public health purposes. As this study encompasses more of the Omicron period, the widening differences in cumulative infection risk between the unvaccinated with prior infection group compared to the previously uninfected but vaccinated group were observed and provide more robust HRs for the Omicron period analysis. As such, we confirm that hybrid immunity, provided by vaccination in individuals with previous COVID-19, was associated with a significantly lower risk of reinfection even during the Omicron period, compared to those with prior infection who remained unvaccinated. This is an important finding as public health decisions regarding future vaccinations are considered. In addition to cross-protective immunity against new VOCs, other advantages of vaccination in previously infected may include longer durability and continued protection against severe disease.

This study supports the recommendation for postinfection vaccination, including providing evidence that the ancestral strain vaccines provided a boosting effect against Omicron in the remotely previously infected, as the prevailing VOC seen in early 2022. Other studies have shown that boosters in patients with prior COVID-19 were associated with additional protection against severe disease (39.3% vaccine effectiveness) [37]. These findings support the notion that a vaccine more concordant with the predominant VOC would provide significant additional protection even in those with prior infection as seen with the recently recommended bivalent vaccine (original and Omicron BA.4/BA.5) [39–42]. Several studies suggest that although protective immunity against infection wanes after 6–14 months [15, 43], T-cell–mediated immunity may continue to provide longer-term immunity but is harder to measure and estimate [10, 19]. The effectiveness of a bivalent mRNA COVID-19 vaccine booster and continued emergence of Omicron sublineages such as XBB.1.5 or other new VOCs warrant continued evaluation for the duration of immunity and timing, especially in persons with remote infections caused by ancestral or other VOCs. The specific vaccine and timing for routine postinfection vaccination will need to be studied with future variant-specific boosters [33, 44].

Strengths of this study include its large size, a study period adequate to provide robust HRs during the Omicron period, and its generalizability to the general public. Limitations include the observational nature of the study, which carries a risk of bias by underidentification of infection among those who did not interact with the medical system for vaccination or testing due to lack of access to care or other socioeconomic barriers. We were able to assess infection only but not hospitalization or other complications. Asymptomatic infections may have gone undetected, as well as infections detected by unreported home antigen testing. The latter factor is the primary reason we selected an end date of 15 March 2022, which corresponds to the wide availability of home antigen tests in the US. Another limitation of the study is that persons with a single mRNA vaccine dose were considered unvaccinated. Other studies have suggested that a single vaccine dose may be as protective as 2 vaccine doses in previously infected individuals [45]. Therefore, categorizing all subjects with only 1 dose of vaccine as unvaccinated could have underestimated the vaccine's ability to boost natural immunity. However, there were only a small number of subjects who met this criterion in the study. Because Omicron emerged late in the study period, all previous infections were remote by this time. Therefore, the protective effect of vaccination seen in the Delta or Omicron period among those with prior infection may not be applicable to those with more recent infection.

In conclusion, vaccination was significantly associated with lower risk of COVID-19 including in those with prior infection, although natural COVID-19 provided robust protection for several months. Vaccination should be encouraged regardless of prior infection as a strategy to boost protective immunity against emerging variants that may share protective epitopes with the vaccine being utilized.

Notes

Patient consent. The design of the study was reviewed and approved by the Mayo Clinic Institutional Review Board (IRB 21-000967) as a minimal-risk study. Patients were excluded if they lacked research authorization per state statutes. The study design did not necessitate patient consent.

Financial support. This work was supported by intramural funding from Mayo Clinic.