Heterologous vaccination interventions to reduce pandemic morbidity and mortality: Modeling the US winter 2020 COVID-19 wave - pnas.org
Significance
Control of the COVID-19 pandemic has been impeded by the slow global uptake of targeted vaccines, emergence of more transmissible variants, and resistance to continuation of nonpharmaceutical interventions. Commonly used vaccines can have nonspecific immune effects, and several have been shown to have beneficial heterologous effects against SARS-CoV-2 infection. However, there is no science-based guidance on effective implementation of such heterologous vaccine interventions (HVIs) to counter the current or future pandemics. We modeled the effect of different HVI strategies on the winter 2020 COVID-19 wave in the United States, finding that targeting both elderly and nonelderly populations and intervening during pandemic growth phases (i.e., effective reproduction number > 1) led to the greatest reduction in morbidity and mortality.
Abstract
COVID-19 remains a stark health threat worldwide, in part because of minimal levels of targeted vaccination outside high-income countries and highly transmissible variants causing infection in vaccinated individuals. Decades of theoretical and experimental data suggest that nonspecific effects of non–COVID-19 vaccines may help bolster population immunological resilience to new pathogens. These routine vaccinations can stimulate heterologous cross-protective effects, which modulate nontargeted infections. For example, immunization with Bacillus Calmette–Guérin, inactivated influenza vaccine, oral polio vaccine, and other vaccines have been associated with some protection from SARS-CoV-2 infection and amelioration of COVID-19 disease. If heterologous vaccine interventions (HVIs) are to be seriously considered by policy makers as bridging or boosting interventions in pandemic settings to augment nonpharmaceutical interventions and specific vaccination efforts, evidence is needed to determine their optimal implementation. Using the COVID-19 International Modeling Consortium mathematical model, we show that logistically realistic HVIs with low (5 to 15%) effectiveness could have reduced COVID-19 cases, hospitalization, and mortality in the United States fall/winter 2020 wave. Similar to other mass drug administration campaigns (e.g., for malaria), HVI impact is highly dependent on both age targeting and intervention timing in relation to incidence, with maximal benefit accruing from implementation across the widest age cohort when the pandemic reproduction number is >1.0. Optimal HVI logistics therefore differ from optimal rollout parameters for specific COVID-19 immunizations. These results may be generalizable beyond COVID-19 and the US to indicate how even minimally effective heterologous immunization campaigns could reduce the burden of future viral pandemics.
On March 16th, 2020, Imperial College London released a landmark report advocating the suppression of SARS-CoV-2 to avoid a pandemic catastrophe (1). Since then, the scientific community has been challenged to create a "bridge period" of reduced COVID-19 morbidity and mortality until safe and effective targeted vaccines are delivered globally (2). Guided by major modeling groups and international and national public health authorities, most countries quickly implemented variably stringent nonpharmaceutical interventions (NPIs) including physical distancing, self-isolation, home working, school closure, and "shielding" of vulnerable populations such as the elderly. Despite ameliorating COVID-19 incidence when applied, these "lockdowns" of regional and national economies also caused severe financial, social, and health repercussions globally (3). In the United States and other countries, resistance to and reversal of NPIs occurred in many jurisdictions, complicating pandemic control and contributing to persistently high COVID-19 incidence.
The rollout of specific COVID-19 vaccines in 2021 led to a temporary reduction of pandemic caseloads in countries with effective vaccine campaigns and ample stocks, but even this has not proven to be the sought-for panacea for epidemiological, logistical, and political reasons. The emergence of virus variants—now dominated by the Omicron and Delta strains—that are more transmissible and pathogenic have reversed many gains achieved to date and have raised questions about the durability of current vaccine efficacy (4). Although a handful of mainly high-income countries have instituted vigorous campaigns that have rapidly provided high coverage, less than 5% of the world's low-income population has received at least one COVID-19–specific vaccination (5), and even in countries with ample vaccine supply, the global phenomenon of multifactorial vaccine hesitancy has led to uneven intranational uptake that has been exploited by the Delta variant. For these reasons, the public health armamentarium against COVID-19 has ample room for adjuncts to both NPIs and COVID-19–specific vaccines.
One as-yet unutilized intervention to potentially prevent SARS-CoV-2 infection and reduce COVID-19 disease is based upon heterologous or nonspecific effects (NSEs) induced by available non–SARS-CoV-2 vaccines (6). The heterologous effect of vaccination refers to the impact that vaccines can have on unrelated infections and diseases. These effects have been noted for almost a hundred years (7), and potential mechanisms include innate and adaptive immune responses. Trained immunity (8⇓⇓⇓⇓⇓–14), increased cytokine production (15⇓–17), viral interference (18), long-lasting type I interferons (19), the antiviral state (20), cross-reactivity (21, 22), and bystander activation (23) are some of the mechanisms proposed.
Some of the best-studied heterologous vaccine actions are from "off-target effects" from the Bacillus Calmette–Guérin (BCG) vaccine (12, 24⇓⇓⇓⇓–29). Epidemiological evidence including several randomized controlled trials (RCTs) have assessed the effect of BCG vaccination on reducing neonatal mortality. In Guinea-Bissau, two RCTs of BCG given to low weight neonates showed reduction in neonatal mortality after BCG, mainly because of fewer cases of neonatal sepsis, respiratory infection, and fever (30, 31). A meta-analysis of three RCTs of BCG-Denmark showed a reduction in mortality rate of 38% at 28 d of life; marked reductions in mortality were also seen within 3 d after vaccination and at 12 mo of age (32). Interestingly, a BCG vaccination prior to an influenza vaccination can boost influenza-specific immunity (33).
Because of the nonspecific benefits of BCG vaccination, a phase III trial called "ACTIVATE-2" assessed whether BCG could protect against COVID-19 in the elderly; prepublication findings suggest a 68% risk reduction for total COVID-19 clinical and microbiological diagnoses (34). A separate study showed that a history of BCG vaccination was associated with a decreased SARS-CoV-2 seroprevalence across a diverse cohort of healthcare workers, and reduced COVID-19 symptoms (35). The magnitude of protective effect against symptomatic disease was similar in both studies: a reported range of 10 to 30% reduction in all respiratory infections in the former and a 34.5% reduction in self-reported diagnosis of COVID-19 in the latter.
Other epidemiological studies have shown NSE benefits from oral polio vaccine (OPV), measles-containing vaccines (MCVs), and several other common immunizations. OPV has been associated with beneficial NSE (20, 36⇓–38) and may become pronounced with subsequent doses (39⇓–41). A systematic review of the associations of BCG, diptheria-tetanus-pertussis, and MCVs with childhood mortality showed that BCG and MCVs reduced overall mortality by more than would be expected through their effects on the diseases they target (42). As with BCG, an RCT of MCV showed a beneficial nonspecific effect on children's survival (43).
Focusing on SARS-CoV-2 transmission, several studies have found that the administration of OPV, Hemophilus influenza type-B, measles mumps rubella (MMR), varicella, hepatitis A/B, pneumococcal conjugate, and inactivated influenza vaccines are associated with decreased SARS-CoV-2 infection rates (44⇓–46). In addition, results from a study in a Dutch hospital showed a 37 to 49% lower risk of SARS-CoV-2 infection in healthcare workers who received the influenza vaccine in the previous flu season, and this finding was also corroborated by a preliminary in vitro study (9). Thus, there is some evidence to support an impact of routine vaccinations on SARS-CoV-2 infection rates, although these ecological studies are prone to bias, do not establish causality, and may be SARS-CoV-2 variant-specific.
Vaccine-mediated heterologous effects could also extend to reducing the severity of COVID-19 disease. There are epidemiological associations between those who have had a past vaccination with BCG, MMR, inactivated influenza vaccine, and recombinant adjuvanted zoster vaccine and reduced mortality and/or reduced COVID-19 severity (35, 45, 47⇓⇓⇓⇓⇓⇓⇓⇓⇓–57), although these additional ecological studies are similarly susceptible to bias. A recent interim analysis of an ongoing clinical trial in Brazil supports this claim, showing that vaccination with MMR reduces the risk of COVID-19 symptoms and need for treatment (58). Given that the COVID-19 pandemic is still a global health emergency (especially in undervaccinated countries) and that the premise of HVI is soundly based in the immunological and epidemiological literature, there is ample reason to consider its potential role as part of a comprehensive package of pandemic control strategies.
The plethora of studies cited can help characterize the hypothetical efficacy of immune system boosting through HVIs to reduce COVID-19 morbidity and mortality. However, estimating the potential real-world effectiveness of such interventions requires their implementation in an environment that can quantify their potential population-level impact in the context of ongoing control measures on viral transmission, health care utilization, and health outcomes. This type of epidemiological projection can be achieved through the use of mathematical models of infectious disease (59⇓⇓⇓⇓⇓–65).
We used the COVID-19 International Modeling (CoMo) Consortium Model (https://comomodel.net), an open-source, age-structured, country-specific, dynamic compartmental model of SARS-CoV-2 transmission and COVID-19 illness, treatment, and mortality, to illustrate how the logistics of implementing a heterologous vaccine intervention (HVI)—in particular, the timing of initiation of such a vaccination campaign in relation to trends in disease incidence and also the age-related population targeting of such a campaign—largely determine the magnitude of their impact. In particular, we instituted an explicitly defined HVI in one of three distinct time frames during the large fall/winter wave of SARS-CoV-2 in the United States (presurge, intrasurge, and postsurge) and across the same total number of individuals in one of three distinct age-targeted population groups (20+ y old, 40+ y old, and 65+ y old).
There are multiple potential applications of heterologous vaccination in this setting, e.g., as a pre–COVID-19 vaccination primer, as a simultaneously delivered or post–COVID-19 vaccination booster (i.e., replacing or delaying the use of a second COVID-specific vaccine dose), or as a solitary "bridging" intervention to reduce or delay COVID-19-related morbidity and mortality until a specific vaccine is available. Here, we explore the last use: that of a solo heterologous vaccination used as a temporizing "bridging" intervention that has only a low level of heterologous effectiveness at reducing viral transmission (here defined as reducing the likelihood of being infected by 5, 10, or 15%) and clinical severity (i.e., reducing the risk of death if infected, again by 5, 10, or 15%). Given the high levels of targeted vaccination now attained in many high-income countries, our results with respect to the prevaccinated US outbreak should be seen as general, model-informed operational guidance that could maximize the beneficial effect of efforts to use common vaccination programs to mitigate and temporize the impact of COVID-19, and possibly future viral pandemics, in the majority of countries worldwide that have not yet received sufficient quantities of COVID-19–specific vaccines to ensure population protection.
Results
Through March 7th, 2021, there were 29,034,160 reported SARS-CoV-2 infections and 524,652 COVID-19 deaths in the US. From March 17th, 2020 (when reporting started), through March 7th, 2021, the US saw an average of 56,405 daily COVID-19 non-intensive care unit (ICU) hospitalizations. Fig. 1A shows the fit of the baseline CoMo Model USA simulation with this reported cumulative mortality and daily hospitalization, while Fig. 1 B and C show the age-stratified mortality and calculated reproductive number (Rt) through the full simulation from February 16th, 2020, to June 30th, 2021. Through March 7th, 2021, the baseline simulation predicts a mean of 28,366,733 reported cases [95% credible interval (CI) 15,137,456 to 45,664,234 across 100 simulations]. The model also projects 525,097 deaths attributed to COVID-19 (95% CI 469,852, 582,443) and a mean of 49,766 non-ICU hospitalizations (27,159, 79,476), including 17,594 (9,821, 27,879) ICU admissions with and without mechanical ventilation. Approximately three-fourths of all deaths occur in the 70+ age group, and 99% occur in individuals 40 y old and older. Rt, the effective reproduction number of the outbreak, rises above 1 from June 26th to July 17th and then again September 22nd through December 3rd, 2020. The baseline model outputs for cases, hospitalizations, and mortality are within 2.5, 1, and 12% of actual publicly available totals for the US COVID-19 outbreak through March 7th, 2021, respectively.*
(A–C) (A) Comparison of CoMo Model output to reported COVID-19–related non-ICU hospitalizations and mortality in the United States of America, February 16th, 2020 through March 7th, 2021 (median prediction ± 95% CI with 0.03 SD Gaussian "noise" applied to 24 key model variables); (B) age-specific mortality and (C) effective reproduction number (Rt) for full simulation February 16th, 2020, through June 30th, 2021.
Introduction of logistically realistic heterologous vaccination campaigns of different efficacy (reducing viral transmission and, independently, disease severity by 5, 10, or 15%) that target three different age-defined populations (fractions of the 20+, 40+, and 65+ age groups, with each cohort containing the same final number of vaccinated adults) at different timepoints in during the US Fall 2020 COVID-19 wave led to different simulated epidemiological and clinical outcomes (Fig. 2 A–D). Across the range of heterologous vaccine effectiveness, "early" interventions (September 1st and October 1st, 2020, prior to or during the onset of the Fall 2020 surge) lead to the largest decline in reported COVID-19 cases (that is, cases that are clinically symptomatic and therefore more likely to be detected; Fig. 2A; Materials and Methods). In contrast, later interventions lead to a larger reduction in both reported and unreported cases, with the greatest effect seen in interventions targeting the broadest targeted age group (20+ y old) starting in December and January (Fig. 2B). Similarly, COVID-19 hospitalization and mortality were most reduced with interventions that begin November 1st and December 1st, 2020, during which time the Rt rose to remain persistently greater than 1 (Fig. 1C), signaling an accelerating pandemic surge. To summarize, across the range of HVI efficacy (5 to 15%), applying the intervention to the broadest age-defined population (in this case, ∼27 million, or 11%, of the US population age 20 and older, compared to 17% of the US population age 40 and older or 50% of the US population age 65 and older, cohorts that also include ~27 million each) during the height of the pandemic surge (that is, with vaccination campaigns commencing December 1st) led to the greatest reductions in total cases, hospitalizations, and mortality compared to the baseline simulation (Table 1). This resulted in a model projection of 477,700 (95% CI ± 22,300) deaths for high- and 510,700 (95% CI ± 26,500) for low-efficacy heterologous vaccines versus a baseline of 591,800 through June 30th, 2021, without an HVI (meaning mortality reductions of 19 ± 4% and 16 ± 5%, respectively). For hospitalizations and reported cases, the corresponding percentage reductions are 24 ± 1% and 22 ± 1% for high-efficacy HVI and 19 ± 3% and 13 ± 3% for low-efficacy HVI, respectively.
(A–D) Results of modeled HVI on reported cases (A), reported and unreported cases (B), hospitalization (C), and reported death (D) varying hypothetical effectiveness at reducing disease transmission (different columns) and clinical severity of disease (different rows), month of campaign initiation (horizontal axis), and population age threshold (11% of 20+ y olds in red, 17% of 40+ y olds in blue, and 50% of 65+ y olds in green).
Heterologous vaccination interventions to reduce pandemic morbidity and mortality: Modeling the US winter 2020 COVID-19 wave - pnas.org
Comments
Post a Comment