SARS-CoV-2 seroprevalence among healthcare workers in a highly vaccinated Japanese medical center from 2020–2023

ABSTRACT

Infection-induced SARS-CoV-2 seroprevalence has been studied worldwide. At Juntendo University Hospital (JUH) in Tokyo, Japan, we have consistently conducted serological studies using the blood residue of healthcare workers (HCWs) at annual health examinations since 2020. In this 2023 study (n = 3,594), N-specific seroprevalence (infection-induced) was examined while univariate and multivariate logistic regression analyses were performed to compute ORs of seroprevalence with respect to basic characteristics of participants. We found that the N-specific seroprevalence in 2023 was 54.1%—a jump from 17.7% in 2022, and 1.6% in 2021—with 37.9% as non-PCR-confirmed asymptomatic infection cases. Those younger than 50 (adjusted OR = 1.62; p < .001) and recipients with 4 doses or less of vaccine had a higher risk to be N-positive, ranging from 1.45 times higher for the participants with 4 doses (p < .001) to 4.31 times higher for the participants with 1 dose (p < .001), compared to those with 5 or more doses. Our findings indicate that robust vaccination programs may have helped alleviate symptoms but consequently caused asymptomatic spread in this hospital, especially among younger HCWs. Although having four doses or less was found to be associated with higher risk of infection, the optimal constitution and intervals for effective booster vaccines warrant further investigations.

KEYWORDS:

• SARS-CoV-2
• COVID-19
• seropositive
• mRNA vaccine
• healthcare worker
• Japan

Introduction

WHO (World Health Organization) declared the end of COVID-19 (coronavirus disease 2019) as a global health emergency in May 2023.¹ Yet, COVID-19 has remained a public health concern with over 114,000 new hospitalizations and 1,300 new admissions to intensive care units (ICUs) globally during the period from 8 January to 4 February 2024.² Moreover, the SARS-CoV-2 (severe acute respiratory syndrome coronavirus 2) virus has continued to mutate, which scientists have been closely monitoring.² Metrics, such as infection rate, have remained in the public interest. Natural infection rates have varied from country to country mainly due to different dominant variants, vaccination campaigns, infection control measures, and public compliance levels. Seroprevalence studies have been an important surveillance tool for investigating infection rates due to the ability to shed light on infection without awareness, especially among highly vaccinated populations.³

In Japan, COVID-19 became a category V infectious disease in May 2023.⁴ Since then, public countermeasures were no longer enforced, including notifiable disease surveillance (complete case reports) of new COVID-19 cases, cluster tracing, traffic restrictions, time-limits for restaurants, required masks in public areas, and self-quarantine for suspicious COVID-19 infections.⁴ Previous sero-epidemiological studies conducted by the National Institute of Infectious Diseases (NIID) of Japan have revealed that infection-induced seroprevalence of Tokyo was estimated to be 3.1% in December 2021, 6.4% in February 2022 and 34.5% in November 2022, indicating that more than half of the population of Tokyo remained uninfected by the end of 2022.^5,6

At Juntendo University Hospital (JUH), a Tokyo-based medical center with highly vaccinated healthcare workers (HCWs) and strict infection control protocols, we have studied seroprevalence using the blood residue from HCWs at annual health examinations since 2020.^7–9 In order to consistently monitor the infection rates among HCWs, we continued our serological study at the 2023 annual health examinations. In addition, because five rounds of mass vaccination programs were organized at our hospital, in this study we aimed to examine the effectiveness of vaccination programs.

Materials and methods

Study participants

In this retrospective cohort study, a total of 3,594 HCWs who participated in the 2023 annual health examination (June 7 through June 19) gave consent for serological analysis. HCWs included medical doctors, nurses, university teachers/researchers/medical students, administrative staff, and co-medical staff (laboratory technicians, radiologists, rehabilitation specialists, pharmacists and nutritionists).

Hospital infection control measures and vaccination campaigns

At our hospital, masks continue to be universally required in all facilities except while eating, a protocol maintained since the outbreak of the COVID-19 pandemic. Temperature checks occur daily at the workplace, with symptoms pointing to COVID-19 requiring further examination; close contacts of confirmed cases are PCR-tested and quarantined (until May 7, 2023). Naturally, our HCWs use public transportation and live among regular citizens in the Tokyo metropolitan area.

A total of five mass vaccination campaigns were organized for HCWs at JUH, from March 2020 to November 2022, all using mRNA COVID-19 vaccines (BNT162b2 and mRNA-1273). The 5th round was administrated using the Omicron BA.4/5 bivalent booster vaccine. Detailed administration periods for the five vaccination campaigns are shown in Figure 1. While the Tokyo metropolitan government offered the 6th round of vaccination (bivalent booster) in the spring of 2023, mainly for seniors, JUH did not organize a mass campaign inside the hospital facility. HCWs were required to report vaccination information to the Department of Safety and Health Promotion of JUH.

Figure 1. Seroprevalence at JUH with respect to the cumulative confirmed COVID-19 cases in the Tokyo metropolitan area.

Dots in the figure indicate seroprevalences at JUH, calculated by the number of N-specific positive HCWs divided by the number of HCWs who participated at annual health examinations with serological analysis.

A total of five mass vaccination campaigns were organized for HCWs at JUH, from March 2020 to November 2022, all using mRNA COVID-19 vaccines (BNT162b2 and mRNA-1273).

Serum testing

The serum samples at the annual health examination were collected. While nucleocapsid (N)-specific and spike (S)-specific antibodies are expected to be generated in individuals infected with SARS-CoV-2, the currently available vaccines in Japan, including the mRNA vaccines used for JUH’s vaccination campaigns, do not generate N-specific antibodies.^7,10,11 Thus, participants with seropositive results with N-specific antibodies can be considered previously infected with SARS-CoV-2. Those with insufficient amounts of serum samples to measure N-specific antibody titers were excluded. Titers of N-specific antibodies were measured using serological assays (Roche Diagnostics GmbH, Mannheim, Germany).

The Elecsys Anti-SARS-CoV-2 (Roche Diagnostics) immunoassay was used with the Cobas e801 analyzer authorized for emergency use by the U.S. Food and Drug Administration under the Emergency Use Authorization to measure SARS-CoV-2 N-specific antibody titers in accordance with the manufacturer’s instructions. The Elecsys Anti-SARS-CoV-2 immunoassay was used to detect total antibodies directed against the N protein and has sensitivity of 100% and specificity of 99.8% (≥14 days after PCR diagnosis).¹² The results are reported as numeric values in the form of a cutoff index (COI; signal sample/cutoff) with qualitative results. A COI ≥ 1.0 is interpreted as a positive test result.

Statistical analyses

Participants’ basic characteristics, vaccination records, and COVID-19 infection history were extracted from the recording system of the Department of Safety and Health Promotion of JUH. The seroprevalence in the HCWs of JUH was analyzed by age, sex, profession, vaccination status, and whether they had a PCR-confirmed COVID-19 infection history. The seroprevalence is presented as crude percentage. Univariate and multivariate logistic regression analyses were performed to compute ORs of seroprevalence with respect to basic characteristics. A two-tailed p < .05 is considered significant.

To assess potential bias due to unmeasured confounders, we calculated the E-values associated with the adjusted ORs (odds ratio) of seropositivity among participants. The E-values estimate how strong the unmeasured confounders would be to explain away the observed association between seropositivity and its factors. The lowest possible E-value is 1, which indicates that no unmeasured confounding can explain away the observed association. The higher the E-value, the stronger the confounder.¹³

Statistical analyses were performed using SPSS Statistics version 29.0 (IBM Corporation, USA) and R software version 4.3.1 (R Foundation for Statistical Computing, Austria).

Ethical approval

This study was approved by the Institutional Review Board (IRB) of Juntendo University Hospital (IRB #M20–0089-M01) and was performed in accordance with the tenets of the Declaration of Helsinki.

Results

A total of 3,594 HCWs participated in this study. The basic characteristics of the participants are shown in Table 1. The median age was 39.3 years old (SD ± 12.3; range 20–87), with 79.7% (2,863/3,594) aged younger than 50 years old. 62.5% were women. By the time of the 2023 health examination (mid-June), 94.9% had received two or more doses of COVID-19 vaccine; 38.0% had received five or more doses. Number of received doses by age group are shown in Supplementary Table S1.

Table 1. N-specific antibody results among participants by characteristic categories (n = 3,594).

	Overall (%)	N-specific antibody positive, n (%)	N-specific antibody negative, n (%)	p-value^a
All	3594 (100)	1944 (54.1)	1650 (45.9)
Age (years)				<.001
20–29	989 (27.5)	606 (61.3)	383 (38.7)
30–39	1064 (29.6)	610 (57.3)	454 (42.7)
40–49	810 (22.5)	431 (53.2)	379 (46.8)
50–59	456 (12.7)	194 (42.5)	262 (57.5)
60 or older	275 (7.7)	103 (37.5)	172 (62.5)
Sex				.635
Male	1348 (37.5)	736 (54.6)	612 (45.4)
Female	2246 (62.5)	1208 (53.8)	1038 (46.2)
Profession				<.001
Medical doctors	920 (25.6)	508 (55.2)	412 (44.8)
Nurses	919 (25.6)	556 (60.5)	363 (39.5)
University teachers/researchers/ medical students	829 (23.1)	443 (53.4)	386 (46.6)
Administrative staff	570 (15.9)	283 (49.6)	287 (50.4)
Co-medical staff^b	356 (9.9)	154 (43.3)	202 (56.7)
No. of vaccine doses received^c				<.001
Unvaccinated	143 (4.0)	88 (61.5)	55 (38.5)
1 dose	39 (1.1)	31 (79.5)	8 (20.5)
2 doses	184 (5.1)	121 (65.8)	63 (34.2)
3 doses	691 (19.2)	453 (65.5)	238 (34.4)
4 doses	1169 (32.5)	648 (55.4)	521 (44.6)
5 doses	1216 (33.8)	548 (45.1)	668 (54.9)
6 doses	151 (4.2)	55 (36.4)	96 (63.6)
unknown	1 (0.0)	0 (0.0)	1 (100.0)
PCR-confirmed COVID-19 history (Yes/No)^d				<.001
No	2259 (62.9)	736 (32.6)	1523 (67.4)
Yes	1335 (37.1)	1208 (90.5)	127 (9.5)
1 time	1295	1169	126
2 times	39	38	1
3 times	1	1	0

^aChi square tests were applied for p values on percentages of N-specific antibody serological results (positive vs. negative) among the different groups of each category.

^bCo-medical staff include laboratory technicians, radiologists, rehabilitation specialists, pharmacists, and nutritionists.

^cGroup of “unknown” was not included in the chi square test.

^dChi square test was conducted among groups “No” and “Yes.” Detailed data of times of PCR-confirmed COVID-19 history were not included for the chi square test.

The seroprevalence at the 2023 annual health examination was 54.1%, with 1,944 (out of 3,594) N-specific positive cases, a jump from 17.7% in 2022.⁷ The seropositive trend from 2020 through 2023 at JUH is shown in Figure 1. Univariate regression analysis found that younger age (p < .001), being nurses (p < .001) and receiving less doses of vaccine (p < .001) are significantly associated with a higher seropositivity rate (Table 2). The multivariate regression model further revealed that compared to HCWs aged 50 or older, younger HCWs had a higher risk for becoming seropositive (adjusted OR = 1.62; 95% CI, 1.36–1.93; p < .001). In addition, compared to those vaccinated with five or more doses, HCWs with four or less doses had a higher risk, ranging from 1.45 times higher for those with four doses (95% CI, 1.23–1.70; p < .001) to 4.31 times higher for those with one dose (95% CI, 1.96–9.47; p < .001). In this model, female participants had a lower risk compared to males (adjusted OR = 0.80; 95% CI, 0.69–0.93; p = .004). Compared to administrative and university staff, nurses (adjusted OR = 1.48; 95% CI, 1.22–1.78; p < .001) had a higher risk to be seropositive; co-medical staff had a lower risk (adjusted OR = 0.75; 95% CI, 0.59–0.95; p = .017) (Table 2).

Table 2. Univariate and multivariate regression analyzes for seropositivity among participants (n = 3,594).

				Univariate regression			Multivariate regression
Parameter	Participants in all, n	N-specific antibody positive, n	N-specific antibody positive, %	OR	[95%CI]	p-value	Adjusted OR	[95%CI]	p-value
All	3594	1944	54.1%
Age
≥50 (reference)	731	297	40.6%
<50	2863	1647	57.5%	1.98	[1.68–2.34]	<.001	1.62	[1.36–1.93]	<.001
Sex
Male (reference)	1348	736	54.6%
Female	2246	1208	53.8%	0.97	[0.85–1.11]	.635	0.80	[0.69–0.93]	.004
Profession
Administrative/University staff ^a (reference)	1399	726	51.9%
Doctors	920	508	55.2%	1.14	[0.97–1.35]	.117	1.18	[0.99–1.40]	.065
Nurses	919	556	60.5%	1.42	[1.20–1.68]	<.001	1.48	[1.22–1.78]	<.001
Co-medical staff ^b	356	154	43.3%	0.71	[0.56–0.89]	.004	0.75	[0.59–0.95]	.017
No. of vaccine doses
5 or more (reference)	1367	603	44.1%
Unvaccinated	144	88	61.1%	1.99	[1.40–2.83]	<.001	1.88	[1.32–2.69]	.001
1 dose	39	31	79.5%	4.91	[2.24–10.76]	<.001	4.31	[1.96–9.47]	<.001
2 doses	184	121	65.8%	2.43	[1.76–3.36]	<.001	2.23	[1.61–3.10]	<.001
3 doses	691	453	65.6%	2.41	[2.00–2.92]	<.001	2.15	[1.76–2.61]	<.001
4 doses	1169	648	55.4%	1.58	[1.35–1.84]	<.001	1.45	[1.23–1.70]	<.001

^aUniversity staff include university teachers, researchers and medical students.

^bCo-medical staff include laboratory technicians, radiologists, rehabilitation specialists, pharmacists and nutritionists.

The E-values in the multivariate logistic model are shown in Supplementary Table S2. The E-value is 1.858 (RR = 1.271) for age and 1.484 (RR = 0.894) for sex. For professions, by using administrative and university staff as reference, the E-value was found to be 1.386 (RR = 1.084) for doctors, 1.724 (RR = 1.214) for nurses, and 1.585 (RR = 0.864) for co-medicals. For vaccine doses, with five or more as reference, we found the E-value was 2.087 (RR = 1.372) for 0 doses, 3.568 (RR = 2.075) for one dose, 2.354 (RR = 1.494) for two doses, and 2.290 (RR = 1.465) for three doses, and 1.698 (RR = 1.203) for four doses.

Among the 3,594 HCW participants, 1,335 (37.1%) had a PCR-positive infection history, including 1,295 testing positive once, 39 testing positive twice and 1 testing positive three times; 2,259 (62.9%) had no record of PCR-confirmed COVID-19 history (Table 1). Among those without PCR-positive records, 736 were found to be N-specific positive, accounting for 37.9% N-positive cases (736 out of 1,944) identified at the 2023 health examination. Among those with PCR-positive records, however, we found 127 participants with N-specific antibody negative results. The 127 participants’ N-specific antibody values and intervals between the latest PCR-positive tests and blood draw at the 2023 annual health examination are listed in Supplementary Table S3.

Discussion

Using specimens from blood residues of HCWs at the 2023 annual health examination, we revealed that the seroprevalence among the participant HCWs of JUH was 54.1%, as of June 2023. Compared to our previous seroprevalence studies conducted with similar methods in mid-2022 (17.7%) and in mid-2021 (1.6%), the seroprevalence increased substantially.^7–9 As we reported previously, before the spread of the Omicron variant in the Tokyo metropolitan area, which started in late 2022, infection at JUH was under control.⁹ This may be attributed to strict infection control measures, such as universal masking and PCR-testing for HCWs with symptoms and close contacts of the confirmed cases, and the dominance of Alpha and Delta variants that were less infectious compared to Omicron. Although no seroprevalence data for Tokyo were available at the same timepoint, we do see the seroprevalence among our HCWs increased proportionally with the cumulative confirmed COVID-19 cases in the Tokyo metropolitan area (Figure 1).

Our studied samples include numbers of HCW participants (38.0%) who received the 5^th or 6^th dose of the bivalent Omicron BA.4/5 mRNA booster vaccine. Previous studies have revealed that receiving the bivalent vaccine proves effective against COVID-19 related hospitalization or symptomatic infection.^14,15 A large Dutch study, aiming to examine the bivalent (original/Omicron BA.1) vaccine’s effectiveness against SARS-CoV-2 infection, concluded a limited added benefit; the Dutch study used information from infection-reported cohorts, which may not fully include infections without awareness.¹⁶ In our study, we didn't have relevant information on hospitalization rate or symptomatic infection, but we found that those with five or six doses of COVID-19 vaccine (bivalent Omicron BA.4/5 mRNA booster), had a lower rate of seropositivity (p < .001). The sensitivity analysis results of our multivariate logistic model also revealed the robust association between vaccine doses and seropositivity among the studied participants (Supplementary Table S2). Additionally, previous studies found that while booster vaccines may improve antibody-mediated immunity and virus neutralization capacity for a certain period of time, they may not substantially augment T-cell responses.^17,18 The optimal constitution and intervals for effective booster vaccines remain to be elucidated.¹⁹

For infection without awareness, we found 37.9% (736 out of 1,944) of those HCWs with N-specific antibody positive results did not have a PCR-confirmed COVID-19 history. What we found in this study is consistent in general with findings of our previous serological study in 2022, which found 48.6% had infection without awareness.⁷ Why the rate of infection without awareness decreased from the previous year is unclear. Nonetheless, for a medical center with regular temperature checks and established PCR testing for HCWs with suspicious symptoms or being identified as close contacts of confirmed COVID-19 cases (until May 7^th, 2023), this finding indicates that infection without awareness is widespread, even in medical centers with strict COVID-19 countermeasures.

Strengths and limitations

This study was the fourth year of our serological studies, beginning in 2020. This continuity provides us the ability to compare the N-specific seroprevalence in 2023 to the previous 3 years, showing seroprevalence change in the same group of HCWs. However, this study has several limitations worth addressing. First, the studied samples are HCWs of a medical center in Tokyo, with relatively younger age and higher vaccination rate compared to the general population of Tokyo. Second, only age, sex, profession and number of vaccine doses were examined; there might be unmeasured confounders associated with seropositivity, such as close contacts with COVID-19 patients and individual hygiene habits outside the hospital facility. Due to the small sample size and limited data available, interpretation of our study results requires caution. Third, we found a number of cases among those with a PCR-confirmed COVID-19 history to be N-negative, which indicates the limitation of COVID-19 antibody detection with a long interval between infection and blood testing. Because the antibodies wane over time, these cases point to a possible lower estimation of the seroprevalence in this study.

Conclusion

Our findings indicate that robust vaccination programs may have helped alleviate symptoms but consequently caused asymptomatic spread in this hospital, especially among younger HCWs. Compared having five or more doses of vaccine, having four doses or less were found to be associated with higher risk of infection, yet the optimal constitution and intervals for effective booster vaccines warrant further investigations.

Author contributions

Prof Naito had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: All authors.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Yan, Naito.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Yan, Nojiri, Urasaki.

Administrative, technical, or material support: Ito, Yamamoto, Horiuchi, Fukuda, Tabe.

Supervision: Hori, Takahashi.

All authors have approved this version to be published and agreed to be accountable for all aspects of the work.

Acknowledgments

The authors thank Kristin Thurlby (Johnson County Community College, Overland Park, KS, USA) for her editorial advice on preparation of this brief report.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

All data used in the study are shown in the figures and tables. There are no more data to disclose.

Supplementary material

Supplemental data for this article can be accessed on the publisher’s website at https://doi.org/10.1080/21645515.2024.2337984.

Additional information

Funding

This research was supported by Japan Agency for Medical Research and Development (AMED) [JP20fk0108472 to TN]. The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation of the manuscript; or decision to submit the manuscript for publication.