The first case of the novel coronavirus disease (COVID-19), caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), was reported in December 2019 in Wuhan city in the People’s Republic of China [1], and the COVID-19 pandemic was declared by the World Health Organization on 11 March 2020 [2]. More than 610,000,000 COVID-19 cases have been confirmed worldwide until 25 September 2022. COVID-19 has caused mortality, with more than 6.5 million deaths worldwide (1.1% of all confirmed cases) [3].

As a prophylactic response to the COVID-19 pandemic, several COVID-19 vaccines have been developed and have received emergency use authorization from the Unites States Food and Drug Administration (FDA) and the European Medicines Agency (EMA) [4–9].

Worldwide, vaccination against COVID-19 commenced and has been implemented at an unprecedented pace including in South Korea, which started COVID-19 vaccination in February 2021. In South Korea, four COVID-19 vaccines (BNT162b2, AZD1222, mRNA-1273, and Ad26.COV2.S) have been administered, and full-vaccination coverage was noted in approximately 80.0% of the population until November 2021 [10].

Vaccine efficacy of more than 80% was reported [11]; however, waning immunity to SARS-CoV-2 after vaccination [12–14] and the appearance of virus variants have both led to a decrease in vaccine effectiveness [15, 16].

In the situation where the vaccine effectiveness for prevention of infection is decreasing and the breakthrough infection are increasing, it is necessary to evaluate whether the additional effectiveness of the vaccine can be expected in addition to the prevention of infection. In other words, it is needed to evaluate whether vaccine is worthwhile even in COVID-19 patients with breakthrough infection, and mitigation of symptoms and decrease in viral load can be useful measure of vaccine effectiveness.

Therefore, this study was conducted to evaluate the impact of COVID-19 vaccines in preventing symptoms and decreasing viral loads with consideration of potential confounders using the real-world data including the overall confirmed COVID-19 cases in a province in Korea.

We compared the demographic characteristics, presenting symptoms and prevalence of comorbidities, and mean Ct values among COVID-19 patients according to the vaccination status (Table 1). Fully vaccinated COVID-19 patients were older, having comorbidities associated with a high risk of severe COVID-19 more, and less likely to have symptoms at the time of confirmation of SARS-CoV-2 infection than the unvaccinated or partially vaccinated groups (all p<0.01). Additionally, higher Ct values for the RdRp gene were observed in the fully vaccinated group compared to those in the unvaccinated and partially vaccinated groups (p<0.01).

We assessed the association between COVID-19 vaccination status and the presence of COVID-19 symptoms, and found a decreased risk of symptomatic COVID-19 in the fully vaccinated group compared to that in the unvaccinated group (RR (95% CI) = 0.93 (0.87–0.99); S2 Table). In contrast, there was an increased risk of symptomatic COVID-19 in the partially vaccinated group compared to the unvaccinated group (RR (95% CI) = 1.05 (1.00–1.11)). Additionally, we evaluated the vaccine effectiveness for each listed symptom of COVID-19 (S2 Table). The risks of cough, sputum, and myalgia were lower in the fully vaccinated group than in the unvaccinated group. In contrast, higher risks of fever and rhinorrhea were observed in the fully vaccinated group (Table 2).

We additionally investigated the impact of COVID-19 vaccination on symptoms in an age-stratified analysis (S3 Table). Though decrease of symptom onset in the fully vaccinated groups younger than 80 years (<40, 40–59, and 60–79 years), the statistical significance was not attained (RR (95% CI) = 0.92 (0.83–1.03) for <40 years; RR (95% CI) = 0.92 (0.82–1.03) for 40–59 years; RR (95% CI) = 0.93 (0.84–1.03) for 60–79 years, respectively).

Furthermore, we compared the least square means of the Ct values of the RdRp and E genes according to the COVID-19 vaccination status after adjusting for potential confounders. The highest Ct value of the RdRp gene was detected in the fully vaccinated group, and a sequential decrease in Ct values was found in the partially vaccinated and unvaccinated groups (mean, 95% CI: unvaccinated, 23.1, 22.7–23.5; partially vaccinated, 24.2, 23.6–24.7; and fully vaccinated, 24.4, 23.8–25.0) (Fig 1A). In post-hoc analysis, we found that the mean Ct value of RdRp gene in unvaccinated group was significantly lower than that in partially vaccinated group and fully vaccinate group. Sequential increase in the Ct value of the E gene was also observed in order of unvaccinated, partially vaccinated, and fully vaccinated patients (mean, 95% CI: unvaccinated, 22.9, 21.7–24.0; partially vaccinated, 23.5, 21.8–25.1; and fully vaccinated, 23.7, 22.1–25.3), however those differences were not statistically significant (Fig 1B).

The impact of COVID-19 vaccination on Ct values of the RdRp and E genes is presented in Table 3. Patients with low Ct value of the RdRp gene (15≤Ct<25) was lower in both partially vaccinated group (RR (95% CI) = 0.96 (0.91–1.02)) and fully vaccinated group (RR (95% CI) = 0.89 (0.83–0.95)) compared to unvaccinated group. When we additionally evaluated the association between vaccination status and the lowest level of Ct value of the RdRp gene (<15), the risk was reduced in both the partially vaccinated group (RR (95% CI) = 0.72 (0.60–0.86)) and fully vaccinated group (RR (95% CI) = 0.71 (0.61–0.84)). Similarly, the COVID-19 partial and full vaccination was associated with decreased viral loads of E gene, though the association was marginally significant for the full vaccination.

Furthermore, the COVID-19 vaccine effectiveness for preventing high viral loads was evaluated in an age-stratified analysis. The risk of the highest viral load of the RdRp gene (Ct <15) was lower in the fully vaccinated group in two age groups (20–39 years and 60–79 years) (S4 Table). The likelihood of the lowest Ct value of the E gene (Ct <15) significantly decreased in the partially vaccinated and fully vaccinated groups aged of 20–39 years (S5 Table).

In this study, we found that full vaccination helped prevent COVID-19 symptoms including cough, sputum, and myalgia in fully vaccinated COVID-19 patients. In addition, full vaccination was associated with a lower viral load in patient aged younger than 40 years.

In previous studies, the proportion of patients with symptoms or severe symptoms (pulse oxygen saturation <90% on room air, or respiratory rate >30 bpm, or features of severe disease such as shock or organ dysfunction) on arrival at the hospital was significantly lower in COVID-19 patients who had received COVID-19 vaccination than in unvaccinated COVID-19 patients [18, 19]. Furthermore, another study reported the symptom-prevention effectiveness of the COVID-19 vaccines, which ranged from a minimum of 49.5% to a maximum of 98.5% according to age groups and days from vaccination to confirmation of SARS-CoV-2 infection [20]. Similarly, we also found that vaccination prevented the occurrence of COVID-19 symptoms. This result may be a less biased estimate of the association between vaccination and presence of symptoms from those reported elsewhere [18–20], because we adjusted for potential confounders and performed age-stratified analyses. However, though we made an effort to control the effects of confounders, we found that fever and rhinorrhea were more prevalent in the vaccinated group compared to unvaccinated group. Therefore, further research is needed to elucidate whether these results are due to insufficient control of confounding factors or whether vaccination actually has a significant association with increased symptoms of fever and rhinorrhea. Although we cannot fully explain why fever and rhinorrhea symptoms are rather prevalent in fully vaccinated patients, it may be derived from the characteristics of fully vaccinated patients rather than the vaccine itself causing symptoms. According to a study reporting differences in symptom patterns according to age in COVID-19 patients, anorexia together with fever was the most common symptom pattern in the patients aged ≥65 years, while the top most symptom pattern in the patients aged of 45–64 years and aged <45 years were nausea together with weakness and dry mouth with sore throat, respectively [21]. In our study, the proportion of elderly was higher in the fully vaccinated group than other groups. Therefore, there would have been many patients with fever in the fully vaccinated group and this may lead to the increased risk of fever in the fully vaccinated group. However, rhinorrhea was discovered as a common symptom even in the patients aged <45 years in the same study, thus further research is needed to explain why the risk of rhinorrhea increased in the fully vaccinated patients.

Though we found little elevated risk of symptom onset in the partially vaccinated group, it is difficult to consider this as clinically meaningful increase. In addition, the symptom group included patients with various spectrums of symptoms, and rhinorrhea was the only symptom with a significantly increased risk in the partially vaccinated group. Therefore, interpretation of the increased risk of symptoms in the partial vaccination group requires caution.

Furthermore, previous studies have reported higher Ct values of RdRp, E, and N genes in vaccinated COVID-19 patients compared to those in unvaccinated patients, which is similar to the results of the present study [22, 23]. However, in these previous studies, potential confounders were not considered when comparing Ct values according to the vaccination status. In contrast, we compared the least square mean of Ct values after adjusting for age, sex, infection route, comorbidity, and nationality. Therefore, a less biased comparison of Ct values according to vaccination status was available in our study. In addition, we found that the vaccine was effective in preventing the highest viral load (Ct <15) in age groups younger than 40 years. However, in contrast to our findings, another study was reported that vaccination had no association with lower viral load [24]. Therefore, additional research is needed to establish the relationship between COVID-19 vaccination and viral load.

Despite the meaningful outcomes of this study, this study has several limitations. First, information on symptoms was telephonically collected only once from the primary epidemiological investigation based on self-report and only symptoms at the time of COVID-19 diagnosis were considered in this study. Therefore, the possibility of misclassification and underestimation of the symptoms might exist. However, this misclassification may have occurred randomly between the comparison groups (unvaccinated, partially vaccinated and fully vaccinated). Random misclassification usually leads the study results toward null, thus reducing the strength of the association between independent variable and dependent variable. Nevertheless, we were still able to find the significant association between the vaccination and the symptom occurrence. Second, the severity of symptoms, such as body temperature or O2 saturation, could not be investigated. Third, though Ct values were measured using RT-PCR kits from different manufacturers, we simply unified all Ct values, as if these values were derived from the same kit. A number of SARS-CoV-2 diagnostic kits based on the RT-PCR method were approved for use by the Ministry of Food and Drug Safety in Korea and there are studies for validation of the performance of these kits [25]. All these kits had the adequate performance with high sensitivity, however differences in Ct values were observed among kits for the same sample. This means that there is a possibility of misclassifying viral loads when simply integrating all Ct values from different kits. However, it is unlikely that a certain comparison group has a high percentage of using a particular RT-PCR kit, but rather it is likely that the use of the kit was randomly assigned among comparison groups. Non-differential misclassification leads to bias toward null, therefore difference in Ct values among kits would lead to the estimation between vaccination status and Ct value toward null value. Fourth, baseline characteristics among comparison groups (unvaccinated, partially vaccinated and fully vaccinated) were quite different. In Korea, vaccination began primarily for inpatients, residents, and workers in nursing hospitals and nursing facilities. Therefore, there were many COVID-19 patients who were older and with underlying diseases in the fully vaccinated group. In addition, there were many patients with a local cluster infection route and the duration from symptom onset to COVID-19 confirmation was short since many confirmed cases in the nursing hospital and nursing facility were discovered through test performed on all residents and workers when COVID-19 cases were found in that facility. Although we considered those differences in characteristics according to vaccination status in the model, it cannot be guaranteed that these differences were sufficiently adjusted.

Despite these limitations, this is the study to evaluate the effectiveness of vaccination on the occurrence of COVID-19 symptoms and viral load after the implementation of COVID-19 vaccination in real-world after controlling for many confounding factors. As this study used large province-based data from more than 9,000 COVID-19 patients, the results of this study may be applicable to the general population.

In conclusion, full vaccination against SARS-CoV-2 could be effective in the era of COVID-19 breakthrough infection increases through a reduction in the symptom occurrence and the viral load. This impact of vaccine on symptoms and viral load in COVID-19 patients could be used in the advocacy to minimize the tendency to vaccine hesitancy or refusal.