Risk factors for severe Covid-19 breakthrough infections: an observational longitudinal study

As of August 31st, 2021, the COVID-19(Coronavirus disease 2019) has affected more than 214 million people and resulted in 3.5 million deaths across almost all countries [1]. In most cases, the symptoms of COVID-19 are mild. However, in certain cases, the disease develops into severe pneumonia and multiple organ failure with a mortality rate of 3.7% [4].The implementation of preventive measures such as infection control practices, social distancing, and border restrictions did not provide a feasible long-term solution [2].Therefore, the emergency use of authorized vaccines was mandatory. The drive to vaccinate large populations or on other terms create a non-natural “herd immunity” is nowadays the main instrument for combating the pandemic and preventing serious disease and death [3].The first vaccination campaigns began in mid-December 2020 in Europe, the Middle East, and North America [4].There are, currently, four main categories of vaccines utilizing a range of established and new vaccine technologies: whole in activated virus (e.g. Sinopharm, Bharat Biotech), Protein Subunit (e.g. Novavax, Sanofi Pasteur), Viral Vector (e.g. Sputnik V, Janssen, AstraZeneca) and RNA based vaccine (e.g. Pfizer/BioNTech, Moderna) [5].Vaccination programs have known successful progress amongst the world despite scientific debate of the variable efficacy reported with different vaccines [6‐11].However, breakthrough infection (post-vaccination infection)still happens after vaccination among fully vaccinated people with changing patterns according to the predominant virus of concern (VOC) [7, 12].In addition, several studies found that not fully vaccinated persons were more likely tobe affected by critical COVID-19 illness leading to hospitalizations and deaths compared with fully vaccinated people [13‐15].Tunisia knew its first epidemic phase of COVID-19 in March 2020 making it mandatory to implement special emergency measures, which help slow down the epidemic spread. However, this epidemic has known four waves resulted in a substantial rise in deaths in the country. As of December 1st, 2021, Tunisia had more than 718,000 reported cases, which resulted in25000 deaths [16].The Tunisian government's COVID-19 citizen vaccination program has begun in early March 2021, and preferentially targets healthcare professionals, individuals aged over 60 years old, individuals with a disease that increases the risk of complication and death, etc. [17]. The vaccination campaign achieved46.6% (21.9–61.5%) fully vaccinated coverage of the population over 15 years on August 31st, 2021 [18].As the vaccination program is gradually expanding its target population, the concerns regarding the ability of the vaccines to decrease the serious cases are increasing.In this paper, we aimed to assess the severity outcomes and to determine its associated factors among vaccinated COVID-19 cases between March and August 2021 in the governorate of Sousse, Tunisia.

To the best of our knowledge, this study is the first to report the above associations in a cohort of vaccinated subjects in Tunisia. In addition, this is a census survey, COVID19 incidence calculated in this study is representative of all the population of the governorate of Sousse. Our study showed that the COVID-19 incidence was higher among people receiving an incomplete vaccination schema compared to fully vaccinated persons consistent with findings in other studies. According to the vaccine type, the highest incidence rate was observed among persons having been vaccinated with the Sputnik V vaccine, and the least incidence was observed with the AstraZeneca vaccine. This finding highlights the discrepancy in the effectiveness of each vaccine on the prevention of transmission potential in the wider populations with different local characteristics and virus variants of COVID-19 infection [21].

Stouten and colleagues [22] found that adenoviral-vector-based vaccines were associated with a higher risk of breakthrough infections, compared to mRNA-based vaccines. In a large real-world cohort of patients on dialysis, Brunelli and colleagues [23] highlighted an inconsistent antibody response in Ad26.COV2.S comparing with BNT162b2, however, no difference was detected in clinical effectiveness over generali the first 6 months postvaccination. However, these findings should be interpreted with caution since the different vaccine types were introduced in different times through the COVID-19 epidemic.

Almost 72.3% of the study population received one dose whereas only 27.7% were fully vaccinated. In addition, the prevalence of severe or critical cases in vaccinated COVID-19 patients occurs in 10.8% of cases, which is in accordance with some of the few studies in the literature [7, 13, 24]. However, The finding for our cohort (10.8%) was much higher than the prevalence (0.5%) observed in previous research between 15 December 2020 and 30 June 2021 in US Veterans Health Administration (VA) healthcare system [25]. This inconsistency may be due to the fact that several factors are interrelated in the occurring of severe COVID-19 disease after vaccination [26]. Yet, these factors could differ and thus influence the rate of serious COVID-19 breakthrough infection from one region to another.

Our current study presents multiple valuable COVID-19-related associations in the group of vaccinated COVID-19 patients during the study period. First, critical COVID-19 disease was noticed in older age groups regardless of vaccination status. Second, the severe or critical cases rate of 35% and 16% was significantly higher among males in the age classes 60–74 years and ≥ 75 years respectively than female participants (17% and 6% respectively) although the percentage of fully vaccinated patients was higher in males than females in these categories of age. Findings in the literature indicated that the male gender is a risk factor for serious COVID-19 disease that is explained by the differences in immunity response, the role of sex hormones, and gender-related behavior [25‐27].

After adjustments, patients with a medical history of cardiovascular diseases had more than two times increased odds to have a serious disease. Immunocompromising conditions (such as hepatological disease, chronic inflammatory diseases, solid or hematological cancers, etc.) could weaken the antibody response to the vaccination [28]. Additionally, Wang et al. noticed between August 4 and October 12, 2021, a higher prevalence of severe breakthrough SARS-CoV-2 delta (B.1.617.2) variant era among patients with cardiovascular and lung disease, type 2 diabetes, history of malignancy, and baseline use of immunosuppressive medications [29].

Patients with cardiovascular diseases were more likely to have a serious disease. On the one hand, the effect of an unsettled immunity system on CVDs' exacerbation was demonstrated in the literature [30, 31]. On the other hand, dysregulated inflammatory response to COVID-19, more precisely neutrophil extracellular traps (NET) formations, induce immune-thrombosis and exacerbate inflammation [32]. These findings may explain the higher risk of developing serious cases of COVID-19 among individuals with CVD compared to individuals without CVD.

We also found the highest self-estimation of adherence to preventive measures was inversely correlated to severe or critical cases. Despite the efficacy of vaccination, preventive measures including the use of face masks, washing, or rubbing hands, and social distancing are crucial to reducing the virus spread whether in terms of speed or the transmitted viral load [24]. Lastly, having an incomplete vaccination schema was strongly associated with severe or critical breakthrough infection. Previous population-based studies showed that unvaccinated adults are far more likely to be hospitalized compared with fully vaccinated adults at least 14 days before the onset of the symptoms. in all adult age groups [7, 33, 34]. However, vaccination was less effective against hospital admission or death among old people compared to younger people [13], especially among those not fully vaccinated [8, 35]. Butt et al. found that the risk of severe outcomes of post-vaccination disease was higher with increasing age, having more than 4 comorbidities, while being vaccinated was associated with strong protection against severe or critical disease [25]. Besides, a previous study revealed that clinical symptoms of the SARS-CoV-2 infection among partially or fully vaccinated cases was significantly lower (odds ratio = 0.26 [95% CI: 0.07–0.94]) than in unvaccinated cases [36].

Dysregulation of immune function like inflammaging and immunosenescence can increase the vulnerability of subjects to COVID-19 and result in inefficient protection after vaccination [37].

Moreover, researchers suggested that the post-vaccination infections might be increasingly registered with the emerging variants of concern (VOCs) [38], which turn out to be less susceptible to neutralizing antibodies [39]. Nevertheless, a previous study suggested that breakthrough infection itself could substantially boost the vaccine-induced antibodies, especially the variant cross-neutralization [40]. There is limited research discussing the efficacy of antibodies cocktails [39], particularly with the high transmissibility of the Delta and Omicron variants [41].

Stopping the spread at the source mitigates opportunities for the virus to mutate by reducing the amount of viral transmission and therefore reducing the circulation of new variants [42]. With respect to the observation that most COVID-19 infections occurred nearly 4 months after launching the vaccination campaign, it was important to notice that during the fourth wave of COVID-19 infections the Delta variant was the main epidemic strain in Tunisia likewise globally [43]. A higher COVID-19 incidence rate was seen during the delta period than during the alpha period. Therefore, our findings may be not only attributable to vaccine effectiveness, but also to the variants of SARS-CoV-2 virus, different social restriction measurements and seasonal effects. Although the vaccination cannot prevent Delta variant infection and morbidity, worldwide data have revealed that the full course vaccinated population has been effectively protected against severe illness and death [44]. Moreover, a retrospective cohort study over 8 months by Tartof and colleagues [45] found that the effectiveness of BNT162b2 vaccine against COVID-19 decreased over time; yet, its effectiveness against hospital admission remained robust among fully vaccinated individuals.

Strengths of our study include evaluation of a regional sample with frequent testing of the population in the governorate of Sousse, availability of all vaccination and testing data in the comprehensive regional COVID- 19 confirmed cases database, subject interviews to confirm socio-demographic and clinical variables. Furthermore, the relatively considerable number of outcomes of interest and characteristics of the population which may be generalizable to other countries and regions.

Though we tried to provide evidence about the post-vaccination infections to improve measures of prevention and control of COVID-19, this study has some limitations.

As reported previously, the laboratory results may show false negatives and fortunately, this is not a frequent technical issue [46]. For testing all samples, several accredited public and private laboratories were authorized which may induce differences in the reliability of tests. Additionally, recall bias could alter the exactness of some information reported by cases. However, the trained investigators have spent sufficient time with each participant to enable them to take back memories and therefore minimize recall bias. Another limitation is that the vaccine may mitigate the symptoms of the SARS-CoV-2 infection. Herein, some asymptomatic people would escape from the COVID-19 screening. Moreover, we included only individuals over 19 years of age in the present analysis because they corresponded to the population initially targeted by the Tunisian vaccination campaign. Vaccination of children and adolescents was made possible at a much later stage (September 2021). Besides, the relatively short follow-up since the vaccination onset may affect the precisions of our estimates and our ability to evaluate the outcomes of breakthrough infections during the epidemic strain of Omicron VOC in the fifth wave. Hence, the generalizability of our findings needs to be validated in other populations including younger ages and other national data resources, and extended to the fifth wave exploring the Omicron VOC with assessing the potential waning after full vaccination. Evaluating serious outcomes of breakthrough infections beyond 8 months will be crucial for updating COVID-19 vaccine policy.

Additionally, comparing the severe outcomes with unvaccinated individuals warrant further investigation. Finally, additional studies should be conducted to quantify the real effectiveness of the combination of various control measures in the Tunisian population data and to provide further guidance on effectiveness differences according to variants of the SARS-CoV-2 virus and the booster shot indication to reduce the severe disease of breakthrough infections.

Our study revealed a relatively low prevalence of severe or critical disease after breakthrough infections among the population of the governorate of Sousse. Furthermore, we investigated the correlation of age, gender, being a healthcare worker, medical history, tobacco use, self-estimation of adherence to preventive measure, vaccination status, and vaccine type with the severity of the COVID-19 disease. Our findings suggested that having cardiovascular disease, poor self-estimation of adherence to preventive measures, and incomplete vaccination schema were important to predict the severity of SARS-CoV-2 infection among a vaccinated Tunisian population. Therefore, boosting immunity for vulnerable patients added to maintaining and promoting preventive measures are not only essential to prevent severe cases of breakthrough infections of COVID-19, but also other influenza-like diseases.