In December 2019, an outbreak of COVID-19, an atypical pneumonia associated to a novel coronavirus (severe acute respiratory syndrome coronavirus 2; SARS-CoV-2) was detected in Wuhan, Hubei province, China.
Epidemiological data showed that SARS-CoV-2 mainly infects and threatens the health of the elderly and subjects with comorbidities, such as diabetes, obesity, cardiovascular, respiratory, renal, and lung diseases [1
]. Due to its high infectiousness, SARS-CoV-2 has spread rapidly throughout the globe, and on 11 March, 2020, the World Health Organization (WHO) officially recognized COVID-19 as a pandemic [2
Italy has been strongly affected by COVID-19 [3
]. The first Italian case was identified on February 21, 2020, at about the same time that the seasonal flu, caused by the influenza virus, reached its annual peak of diffusion [4
]. The diseases caused by influenza virus and SARS-CoV-2 both share a similar route of transmission (i.e., through aerosolized or respiratory droplets) and some respiratory and systemic symptoms, but they strongly differ in terms of rates of severe and fatal cases [5
], and peculiarly, of the age groups predominantly affected. Indeed, influenza preferentially affects children and young adults [6
], whereas the symptomatic SARS-CoV-2 infection rate increases with age, from about 50–100 cases/100,000 in children and subjects younger than 20 to about 900 cases/100,000 in subjects over 80 years old [7
]. One possible explanation for the low susceptibility in children and young people to the SARS-CoV-2 infection is a more effective and reactive immune system, boosted by the exposure to common pediatric viral agents or to viral antigens contained in the numerous anti-viral vaccines (measles, mumps, rubella, varicella, hepatitis B, hepatitis A, rotavirus, papilloma virus) administered early in life. Indeed, vaccines may induce positive “non-specific” immunotherapeutic mechanisms improving the host’s response to other pathogens, through a process named “trained immunity” [9
Annual vaccination, a major component of public health programs, is the most efficacious and cost-effective intervention to prevent seasonal influenza [10
]. In addition, during the coronavirus SARS-CoV-1 outbreak which occurred in China at the end of 2002, the WHO recommended increasing the coverage rate of influenza vaccination, particularly in frail groups (e.g., elderly and disabled people) [11
As neither a specific vaccine against SARS-CoV-2 nor anti-COVID-19 drugs are available yet, and since several pieces of evidence suggest a real risk of a second wave of outbreaks [12
], it may be relevant to assess whether influenza vaccination could mitigate a new COVID-19 outbreak. As a first investigational approach, in the present study we evaluated whether the different influenza vaccination coverage rates reached in people aged 65 and over in each Italian region are associated with a different spread of the COVID-19 outbreak.
Using aggregated data of Italian regions and AP obtained from official national websites, in this ecological study, we found an inverse association between the extent of influenza vaccination coverage rate and the seroprevalence of SARS-CoV-2, the prevalence of patients hospitalized, admitted to intensive care units, or the number of deaths attributable to COVID-19. The strength of these associations, observed at univariate analyses, markedly increased after adjustment for potential confounders, such as the percentage of health expenditures with respect to the regional Gross Domestic Product, the mean seasonal temperature, the delay in applying the lockdown, the Italy–China import–export commerce, and the cardiovascular mortality rate. Indeed, compared with univariate analyses, the adjustment for confounding factors increased the model’s R2
by about 159% considering SARS-CoV-2 seroprevalence, ~82% considering patients hospitalized with symptoms, ~94% considering patients hospitalized in intensive care units, and ~136% considering the number of deaths attributable to COVID-19 (Table S2
). Most importantly, based on adjusted regression coefficients (Table 2
), it can be estimated that a 1% increase in the vaccination coverage rate among subjects aged 65 and over (i.e., about 140,000 doses of vaccine throughout Italy) would have resulted, in the entire Italian population (60.36 million inhabitants), in a reduction of 78,560 seropositive subjects, 2512 hospitalized patients with symptoms, 353 patients hospitalized in intensive care, and 1989 deaths.
To the best of our knowledge, only a few studies have examined the relationship between influenza vaccination and COVID-19 outcomes [32
]. In an ecological study carried out at a county-level in the American elderly population, an inverse association between influenza vaccination coverage rate and deaths attributable to COVID-19 was reported [33
]. A similar association was found in an Italian ecological study, although through performing an unadjusted analysis [34
]. In our study, we corroborate and extend these observations by showing, in the Italian population, a potential protective role of influenza vaccination on COVID-19 mortality also after adjusting the data for a variety of potential confounders. In addition, we found that influenza vaccination coverage rate is also independently associated with the SARS-CoV-2 seroprevalence and the occurrence of non-fatal clinical expressions of COVID-19, indexed by the rates of hospital admissions and intensive care unit admissions. Our data are in line with a recent Brazilian study [32
] reporting a significantly lower risk of hospitalizations in intensive care treatment, of invasive respiratory support, and of death in COVID-19 patients who had recently received influenza vaccination than in those who had not.
Although our findings do not allow for assumptions to be made about the mechanisms underlying the putative protective role of influenza vaccination on COVID-19 outcomes, previous immunological and epidemiological studies [35
], as well as mathematical models [40
], support the concept that vaccination against one microorganism may affect the host’s response to other infectious agents. For example, vaccination against mycobacterium tuberculosis with the bacillus Calmette–Guérin vaccine significantly increases the secretion of IL-1B (a pro-inflammatory cytokine), which plays a recognized role in antiviral immunity [43
Alternatively, preventing influenza through vaccination might reduce the risk of respiratory superinfection with SARS-CoV-2. However, a recent report about the rates of co-infection between SARS-CoV-2 and other respiratory pathogens suggests that influenza virus and SARS-CoV-2 co-infection is extremely rare (0.9%) [44
]. Yet, this coinfection might determine an enhanced COVID-19 severity, as was found in more than 20% of patients dead from SARS-COV-2 in Northeastern Iran [45
]. Finally, people who agree to undertake influenza vaccination may merely have a more proactive attitude towards prevention, including a respect of social distancing, proper use of face masks, and the use of other personal protection equipment (PPE) during the pandemic compared to non-vaccinated people, which could have determined less COVID-19 outcomes in groups with a high rate of influenza vaccination coverage.
Our incidental observation at univariate analyses of a direct correlation between regional influenza vaccination coverage rate and mortality from cardiovascular diseases is counterintuitive and should be interpreted with caution. Indeed, a variety of studies have indicated that influenza vaccination is an effective intervention to reduce cardiovascular risk [46
], and guidelines endorse its administration, mainly in secondary cardiovascular prevention or other conditions of frailty [49
]. Therefore, we believe that one possible explanation for the observed association may be “reverse causality” (i.e., more people are vaccinated against seasonal influenza in regions with a higher prevalence of frail individuals, as revealed by a higher cardiovascular mortality rate).
Not surprisingly, in this study, the regional mortality from cardiovascular diseases was positively and independently associated with the regional rate of hospitalizations for COVID-19, a finding in line with numerous reports showing cardiovascular disease as a predisposing factor to symptomatic and more severe clinical expressions of COVID-19 [25
This study has several strengths. First, we used aggregated data from official national web sites, which allows for generalizing our findings to the whole Italian population. Second, having used national level data, our analyses have eluded the effect of potential confounding transnational differences (e.g., cultural norms, medical care standards, health system policy, government response in duration of quarantine or timeliness of the start of the blockade, and homogeneity of data collected), which, in absence of a statistical adjustment [50
], might provide spurious associations.
This study also has some limitations. First, as the 2019–2020 regional influenza vaccination coverage rate was not available at the time of our study, our analyses are based on an extrapolation of data of the last five years (see Figure S1
). Second, despite our attempt to control for the effects of some of the most important confounders identified in the literature, we cannot exclude an influence on the results of other confounders that have not been identified yet. Third, regional rather than individual data have been used. Therefore, we cannot exclude the ecological fallacy (i.e., the fact that the observed associations might not persist at an individual level) [51
]. Fourth, Italian data cannot be automatically extrapolated to other countries.
Pending corroboration through experimental clinical studies with proper design, the results of our ecological analyses might have major public health implications. Indeed, they suggest that the influenza vaccination coverage rate of the population, on top of the currently recommended measures to limit contagion (social distancing, confinement, use of PPE), might contribute to attenuate the SARS-CoV-2 pandemic.