SARS-CoV-2 and Its Variants in Thrice-Infected Health Workers: A Case Series from an Italian University Hospital

Background: We described a SARS-CoV-2 thrice-infected case series in health workers (HW) to evaluate patient and virus variants and lineages and collect information on variables associated with multiple infections. Methods: A retrospective analysis of clinical and laboratory characteristics of SARS-CoV-2 thrice-infected individuals was carried out in Verona University Hospital, concurrent with the ORCHESTRA project. Variant analysis was conducted on a subset of available specimens. Results: Twelve HW out of 7368 were thrice infected (0.16%). Symptomatic infections were reported in 63.6%, 54.5% and 72.7% of the first, second and third infections, respectively. Nine subjects were fully vaccinated at the time of the third infection, and five had an additional booster dose. The mean time to second infection was 349.6 days (95% CI, 138–443); the mean interval between the second and third infection was 223.5 days (95% CI, 108–530) (p = 0.032). In three cases, the second and third infections were caused by the Omicron variant, but different lineages were detected when the second vs third infections were sequenced. Conclusions: This case series confirms evidence of multiple reinfections with SARS-CoV-2, even from the same variant, in vaccinated HW. These results reinforce the need for continued infection-specific prevention measures in previously infected and reinfected HW.


Introduction
SARS-CoV-2 reinfections were reported in mid-2020, raising concerns about natural immunity [1]. The onset of SARS-CoV-2 reinfection represents an obstacle in handling the pandemic since it defies the herd immunity concept and control measures [2]. It is reported that SARS-CoV-2 can reinfect fully vaccinated individuals. The frequency of reinfection was not determined among unvaccinated, partially vaccinated, and fully vaccinated individuals, even though the vaccination reduces the severity of infection [3]. It is noteworthy that a key role in reinfection is played by SARS-CoV-2 genome mutations, thus inducing the appearance of new variants with different clinical characteristics [4]. A deeper understanding of viral and immunologic features of SARS-CoV-2 reinfections may help define reliable correlates of immunity [5].
In this report, we evaluated individual clinical variables, virus variants, and lineages in a series of thrice-infected health workers (HW).

Results
Twelve HW thrice SARS-CoV-2 infected (8 males and 4 females) were detected, with a mean (±SD) age of 44 y (±9. 4). Clinical data were available for 11 HW: two had allergies, one had hypertension. BMI was in the healthy weight range for all HW. Moreover, 14/33 infections were occupational, 8 originated from household contacts. The source was unknown in 11.
At the time of 33 infections, 10 HW were not vaccinated, 10 fully vaccinated, 7 up to date boosters, and 6 partially vaccinated. As regards the vaccine type, 11/12 HW received BNT162b2, while 1 HW received only two doses of mRNA-1273. 30% of reinfections were considered breakthrough infections.
The mean interval was 349.6 days (95% CI, 138-443) between the first and second infections, while the mean interval between the second and third infection was 223.5 days (95% CI, 108-530) (p = 0.032) Table 1 shows HW characteristics and infection details, including the variant analysis for each individual. Occupational Relative/friend Occupational * Not laboratory identified but assumed based on the loco-regional epidemiological data on the dominant variant [8]. ** Samples previously tested positive were analysed to assess the SARS-CoV-2 variant by Novaplex™ SARS-CoV-2 Variants VII Assay (Seegene, Seoul, South Korea), enabling distinction between Alfa, Beta/Gamma, Delta, and Omicron BA.1 and BA.2 variants of concern, following manufacturer's instructions. † SARS-CoV-2 subvariants obtained by RNA sequencing analysis. Figure 1 illustrates a radial phylogenetic tree of sequenced SARS-CoV-2 strains for the analysed patients, according to the designated clades of the virus. The tree was generated by the Nextclade site comparing the two different Omicron lineages of the second and third infection of patients 3, 4, and 9, displaying the phylogenetic distance between the two samples. In all cases, an Omicron variant was detected. Figure 1 illustrates a radial phylogenetic tree of sequenced SARS-CoV-2 strains for the analysed patients, according to the designated clades of the virus. The tree was generated by the Nextclade site comparing the two different Omicron lineages of the second and third infection of patients 3, 4, and 9, displaying the phylogenetic distance between the two samples. In all cases, an Omicron variant was detected. Since the Novaplex kit is not able to differentiate the Omicron variant BA.4 and BA.5 with respect to the BA.1 and BA.2, we performed sequencing analysis by NGS procedure in these samples. The results indicated that the lineages were different. In fact, all second infections were classified in the 21K Omicron variant, during the third infection in the 22B Omicron variant (Figure 1, Table 1) in the BA.5 lineage. Table 2 displays symptom categories (no symptoms, minor, major, hospitalisation) and vaccination status in first, second, and third infections, while Figures 2 and 3 detail the type of symptoms. The median value (IQ25-75) of symptoms duration among 1st, 2nd, and 3rd infections was 4 days (0-6), 0 (0-4) and 3 (1, 5-3,5), respectively.  Since the Novaplex kit is not able to differentiate the Omicron variant BA.4 and BA.5 with respect to the BA.1 and BA.2, we performed sequencing analysis by NGS procedure in these samples. The results indicated that the lineages were different. In fact, all second infections were classified in the 21K Omicron variant, during the third infection in the 22B Omicron variant (Figure 1, Table 1) in the BA.5 lineage. Table 2 displays symptom categories (no symptoms, minor, major, hospitalisation) and vaccination status in first, second, and third infections, while Figures 2 and 3 detail the type of symptoms. The median value (IQ25-75) of symptoms duration among 1st, 2nd, and 3rd infections was 4 days (0-6), 0 (0-4) and 3 (1, 5-3,5), respectively.

Discussion
Very few data are available on SARS-CoV-2 thrice infected. To our knowledge, only another study with a similar number of cases of HW infected more than twice is available [9] in a tenfold larger population. Our data could be linked to HW periodical SARS-CoV-2 screening, but a similar outcome might occur in the general population.
As reported by Swift et al., multiple infections can also occur in young, immunocompetent and vaccinated individuals, and comorbidities do not seem to play a key role. Indeed, among the three subjects who reported comorbidities in our study (25%), two had mild allergic respiratory diseases, and one had arterial hypertension on drug therapy. Neither in this study nor in Swift's was affected by immunosuppression. It is therefore probable that one or multiple previous infections, as well as vaccination and comorbidities, influence the disease's severity rather than the risk of infection, in particular after the onset of variants of concern (VOC).
Even a young age does not seem to be protective against the risk of multiple reinfections. Both our study and Swift's showed a low median age (46 and 27, respectively) [9].
Regarding the interval between the infections, we found that the second and third infections had a shorter lag time than that between the first and second infections. These results align with those highlighted by Swift et al., and they seem to suggest that VOC also have a major impact on this aspect.
Our data show that screening testing maintains a primary role in the prevention of the infection spreading, especially in high-risk categories and previously infected subjects. Indeed, fourteen out of 33 (42.4%) infections were detected through periodic testing. Similar results (9/33; 27.3%) were found by Swift et al. (also including pre-and post-travel screening) [9].
Although the small number of infections does enable definitive conclusions, this study has some strengths. In many samples, it was possible to identify the SARS-CoV-2 variant related to infection and Ct values. This information increased the specificity of our definition of reinfection. Moreover, clinical data and vaccination status were collected, enabling a better description of infections.

Conclusions
This study shows multiple SARS-CoV-2 reinfections also in vaccinated HW; interestingly, three patients showed Omicron variant both in the second and third infection, but different lineages were detected.
The continuous infection-specific prevention measures and targeted screening programmes with swabs still represent valuable infection control procedures, especially in at-risk populations such as HW.