What this is
- This research investigates SARS-CoV-2 infection rates among vaccinated healthcare workers in Manaus, Brazil.
- The study occurred during the emergence of the Gamma variant from March to May 2021.
- It involved 771 healthcare workers, assessing their infection rates and vaccination status.
- Findings indicate a low attack rate of COVID-19 in a highly vaccinated cohort.
Essence
- The COVID-19 attack rate among vaccinated healthcare workers was low at 5.2 per 1000 HCW-weeks. Most infections occurred in fully vaccinated individuals, who cleared the virus more rapidly than unvaccinated counterparts.
Key takeaways
- The overall COVID-19 attack rate was 5.2 per 1000 HCW-weeks, with 16 confirmed infections. Among these, 12 were fully vaccinated, indicating that vaccination did not eliminate infection but may aid in faster viral clearance.
- Infected vaccinated healthcare workers had a median Ct value of 32.5, suggesting lower viral loads compared to unvaccinated individuals, who had a median Ct value of 26.5. This implies that vaccinated individuals may experience milder infections.
- The study found that healthcare workers with high-titer antibody responses had a 56% lower risk of infection, although this finding was not statistically significant. This suggests a potential correlation between antibody levels and infection risk.
Caveats
- The low attack rate may limit the statistical power to detect significant associations between exposure and infection risk. This could affect the generalizability of the findings.
- Self-reported data on prior infections and PPE use may introduce recall bias, potentially skewing results regarding infection risk and protective measures.
- The study's follow-up duration of three weeks may not capture longer-term infection dynamics, particularly as variants evolve.
Simplified
METHODS
Study Setting
Manaus, the capital of Amazonas state, is the most populous city in the Brazilian Amazon region, estimated to have 2.06 million inhabitants [30]. From 1 January–31 March 2021, Manaus reported 77 967 new confirmed COVID-19 cases and 8609 COVID-19–related hospitalizations [31]. We enrolled HCWs from 2 major public hospitals in Manaus: (1) Hospital Pronto Socorro 28 de Agosto (HPS 28), a 379-bed hospital for adult and pediatric patients, including 64 COVID-19 intensive care unit (ICU) beds, located in the south-central region of Manaus and with approximately 2800 HCWs, and (2) Hospital Pronto Socorro Platão Araújo (HPS Platão), a 189-bed hospital for adult patients only, including 33 ICU beds and 22 step-down unit beds, located in the eastern region of Manaus and with approximately 1900 HCWs. Throughout January–March 2021, these hospitals were dedicated to COVID-19 admissions, suspending other services. Despite having established infection prevention and control (IPC) teams, both hospitals faced overcapacity, necessitating patient airlifting to other states for care.
Study Design and Population
We conducted a prospective, observational cohort study of HCWs from diverse cadres. Healthcare workers from both hospitals were enrolled from 31 March–31 May 2021, and were followed up for 3 weeks, for a total of 4 visits. A nasal swab was collected at enrollment and weekly regardless of symptoms, and a blood specimen was obtained at enrollment. Healthcare workers were eligible for inclusion if they had direct contact with patients or were in contact with patients' surroundings (eg, hospital cleaners), were 18 years of age or older, and had a way to be contacted by phone. To ensure consistent exposure to the hospital environment, we enrolled only HCWs who reported working at least 20 hours per week. Healthcare workers were excluded if they were an employee of the hospital but not on active duty since December 2020, worked fewer than 20 hours per week at the study hospital, or refused collection of nasal swab or blood sample at enrollment.
Data Collection and Definitions
Study staff were trained on specimen and data collection and recruited HCWs at the hospital 5 days per week, spanning multiple shifts to accommodate HCWs across all work schedules. Data were collected using a standardized tool and included demographics, occupation, chronic conditions, occupation risks, COVID-19 community exposures, COVID-19 vaccination, history of COVID-19, and clinical symptoms. Weekly follow-ups included a shorter questionnaire on exposures, clinical symptoms, personal protective equipment (PPE) use, and COVID-19 vaccination updates. Documentation of COVID-19 vaccination status, including formulation, dose, and dates, was obtained from participant immunization cards or the state COVID-19 immunization registry [32].
We collected data on 12 aerosol-generating procedures (AGPs), categorizing them by risk level based on WHO guidelines (). We assessed HCWs’ involvement in performing, assisting, or being present during AGPs. Vaccination status was classified as fully vaccinated (2 doses ≥14 days before specimen collection), partially vaccinated (1 dose ≥14 days before specimen collection), or unvaccinated (no documented doses). Infection despite vaccination was defined as confirmed SARS-CoV-2 infection by real-time reverse-transcriptase polymerase chain reaction (RT-PCR) in fully vaccinated HCWs. Prior COVID-19 infection relied on self-reporting by HCWs. Supplementary Table 1
Laboratory Methods
Self-collected nasal swabs were placed in viral transport medium, stored at 2°C–4°C, and transported in coolers within 48 hours after collection to Laboratório Central do Amazonas (LACEN-AM) for real-time RT-PCR for SARS-CoV-2 using AllplexTM 2019-nCoV Assay [33]. The RT-PCR results were expressed as the cycle threshold (Ct) for the gene encoding the nucleocapsid protein (N gene), as previously described [34]. All samples with sufficiently low Ct values (Ct <30) from RT-PCR testing were defined as a case and underwent whole-genome sequencing using a COVIDSeq library preparation kit (Illumina) at Fiocruz-Amazonas, as previously described [2]. The Ct values from the qualitative PCR were used as a proxy measure for viral load.
Whole blood (4 mL) was placed in EDTA-coated tubes and plasma was tested using previously validated in-house enzyme-linked immunosorbent assay (ELISA) for SARS-CoV-2 nucleocapsid (NC) immunoglobulin G (IgG) antibody titers (residues 1–410; GenBank: QHD43432.2↗) [7], and commercially available ELISA for SARS-CoV-2 S1 IgG titers (EUROIMMUN Anti-SARS-CoV-2 S1 Curve ELISA) [35]. An anti–SARS-CoV-2 NC IgG antibody reactivity index (RI) was expressed as the ratio between optical density of the patient sample and the negative control. All samples with an RI of 1.5 or greater were considered positive. For anti–SARS-CoV-2 S1 IgG, we used the manufacturer’s cutoff of more than 35.2 binding antibody units/mL (BAU/mL) to determine reactivity. Reactive samples were further classified as low-titer if BAU/mL was 35.3–260.03 and high-titer if greater than 260.03 based on the distribution of the serologic titers among our HCW cohort. This classification is consistent with data from ChAdOx1 nCoV-19 clinical trial that found a titer of 264 BAU/mL or greater to be correlated with protection against SARS-CoV-2 infection [36].
Statistical Analyses
We summarized HCW characteristics by hospital using frequencies for categorical variables and medians with ranges for continuous variables. Attack rates were calculated as the number of new infections per 1000 susceptible HCW-weeks. Susceptibility to infection was defined as testing negative for SARS CoV-2 at enrollment or in previous study visits. A Fisher's exact test was conducted to assess the relationship between vaccination and persistent positive tests. Logistic regression with generalized estimating equations (GEEs) was used to calculate odds ratios (ORs) across exposures for our primary outcome of positive RT-PCR–confirmed SARS-CoV-2 infection. This nested model accounts for clustering at the facility level and for repeated measures for each HCW across the study period [37].
All analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC). This project was reviewed and approved by the Institutional Review Board from Universidade de São Paulo and the Brazilian National Ethics Committee. This activity was reviewed by the Centers for Disease Control and Prevention (CDC), deemed non-research, public health investigation, and was conducted consistent with applicable federal law and CDC policy. Written informed consent was obtained from all participants prior to enrollment.
RESULTS
We included 771 HCWs, 383 from HPS 28 and 388 from HPS Platão. The median age was 40 years (IQR: 31–48 years), 600 (78%) were female, 449 (58%) were either a registered nurse or nurse assistant, and 660 (86%) reported no comorbidities. Of the 466 HCWs who reported COVID-19 prior to study enrollment, 84 (18%) reported onset within 90 days of recruitment. COVID-19 vaccination coverage for at least 1 dose was 84% (644/771), with the majority receiving CoronaVac (98%; 628/644) (Table 1). Of the 771 HCWs, 38% reported caring for patients with suspected or confirmed COVID-19 in the 2 weeks prior to study enrollment or thereafter. Among those, 91% reported wearing a cloth mask most of the time or always, 383 (34%) reported wearing medical masks, and 247 (22%) reported wearing N95 respirators (Supplementary Figure 1). Overall, 70% of HCWs did not report exposure to potential AGPs, while 18% encountered high-risk exposures, 8% medium-risk, and 4% low-risk exposures during the study (Supplementary Table 2).
SARS-CoV-2 infection was confirmed in 16 (2%) participants during the follow-up period, and none were hospitalized. The first case occurred on 12 April 2021, during the later stages of the epidemic curve, at a point when daily cases of COVID-19 in Manaus were declining (Figure 1). The median age among infected HCWs was 43.5 years, with 6 (38%) reporting a prior infection. Most of the infected HCWs provided regular hands-on medical care (9 HCWs). Among those, 4 reported providing direct care for a patient with suspected or confirmed COVID-19. Of those 4 HCWs, 1 reported no PPE use in the week before testing positive and the remaining 3 reported only using a cloth mask with no other PPE.
The majority of infected HCWs (12/16; 75%) were fully vaccinated, while 3 infections occurred among unvaccinated HCWs and 1 among a partially vaccinated HCW. COVID-19 symptoms were reported by 58% of vaccinated HCWs compared with 75% of unvaccinated or partially vaccinated HCWs. All 3 HCWs who reported fever were unvaccinated. The median time from completion of the 2-dose vaccine schedule to infection among the 12 infections in vaccinated HCWs was 68.5 days (range: 48–106 days). The median Ct among these cases was 32.5 (range: 24–36) compared with 26.5 (range: 20–32) among unvaccinated HCWs (Figure 2). Two out of the 3 unvaccinated HCWs remained positive after 1 week, while none of the infected vaccinated HCWs tested positive on the 1-week follow-up visit (P = .03).
The overall COVID-19 attack rate per 1000 susceptible HCW-weeks was 5.2: 5.9 for males and 5.0 for females. By job category, the attack rate was highest among environmental cleaning staff and nurse assistants (7.7 and 6.9 per 1000 HCW-weeks, respectively) (Table 2). The attack rate was higher for those aged 60 years and older compared with the younger age group (9.3 vs 5.1 per 1000 HCW-weeks) and for those without prior COVID-19 (7.5 per 1000 HCW-weeks). Healthcare workers with known direct contact with patients with confirmed or suspected COVID-19 had a relatively lower attack rate than HCWs who reported no direct contact with patients with COVID-19 (3.9 vs 6.5/1000 HCW-weeks). The attack rate among participants who received at least 1 dose of a COVID-19 vaccine was 5.1 per 1000 HCW-weeks compared with an attack rate of 5.9 per 1000 HCW-weeks among unvaccinated HCWs (Table 2). Vaccination was not statistically associated with a decreased risk of infection (OR: .89; 95% CI: .25–3.15). Healthcare workers with high-titer reactivity to the S1 or spike protein were estimated to have a 56% lower risk of SARS-CoV-2 infection, and those with low-titer reactivity had a 33% lower risk compared with those with nonreactive serology (Table 2). However, these findings were not statistically significant. Due to high Ct values, only 6 specimens (2 from unvaccinated and 4 from vaccinated HCWs) were sequenced, and all of them were identified as Gamma (P.1).
![Epidemiologic curve of daily COVID-19 cases in Manaus, Brazil, from March 2020 to May 2021. Abbreviation: COVID-19, coronavirus disease 2019. *The blue dot may represent >1 study case. Laboratory-confirmed COVID-19 data was extracted from the Fundação de Vigilância em Saúde do Amazonas (FVS-AM) website.[]. https://www.fvs.am.gov.br [8]](https://europepmc.org/articles/PMC12497958/bin/ciaf339f1.jpg.jpg "Click to view full size")
Epidemiologic curve of daily COVID-19 cases in Manaus, Brazil, from March 2020 to May 2021. Abbreviation: COVID-19, coronavirus disease 2019. *The blue dot may represent >1 study case. Laboratory-confirmed COVID-19 data was extracted from the Fundação de Vigilância em Saúde do Amazonas (FVS-AM) website.[]. https://www.fvs.am.gov.br [8]
SARS-CoV-2 Ct values for infected HCWs (n = 16) and time to infection after full vaccination (2 vaccine doses >14 days prior to positive specimen) (n = 12)—Manaus, Brazil, March–May 2021. Unvaccinated HCWs are displayed on the-axis time point 0 days since they were unvaccinated. *Had persistent viral shedding (defined as testing positive for >1 weekly visit). **Partially vaccinated (1 dose). Abbreviations: Ct, cycle threshold; HCW, healthcare worker; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2. y
| HPS Platão (n = 388) | HPS 28 (n = 383) | Total, No. (%)(N = 771) | |
|---|---|---|---|
| Age, median [IQR], y | 41 [32–47] | 40 [30–48] | 40 [31–48] |
| Sex | |||
| Male | 67 | 104 | 171 (22.18) |
| Female | 321 | 279 | 600 (77.82) |
| Occupation | |||
| Nurse assistant | 198 | 133 | 331 (42.9) |
| Registered nurse | 66 | 52 | 118 (15.3) |
| Security | 10 | 59 | 69 (8.9) |
| Environmental cleaning staff | 30 | 36 | 66 (8.6) |
| Administrative staff | 12 | 15 | 27 (3.5) |
| Physical therapist | 12 | 21 | 33 (4.28) |
| Physician | 8 | 16 | 24 (3.1) |
| Other | 52 | 51 | 118 (15.3) |
| No. of healthcare facilities working at | |||
| 1 | 289 | 303 | 592 (76.8) |
| 2 | 96 | 78 | 174 (22.6) |
| 3 | 3 | 2 | 5 (0.65) |
| No. of hours with patient face-to-face contact during typical work week (1–80 hours), median [IQR] | 20 [10–30] | 30 [1–40] | 25 [10–39] |
| HCW regularly provides hands-on medical care | |||
| Yes | 241 | 184 | 425 (55.1) |
| No | 147 | 199 | 346 (44.9) |
| Smoking status | |||
| Every day | 11 | 20 | 31 (4) |
| Somedays | 16 | 21 | 37 (4.8) |
| Not at all | 361 | 342 | 703 (91.2) |
| Participated in physical activity in the last 30 d | |||
| Yes | 132 | 146 | 278 (36.1) |
| No | 256 | 237 | 493 (63.9) |
| Comorbidities | |||
| Yes | 56 | 55 | 111 (14.4) |
| No | 332 | 328 | 660 (85.6) |
| Baseline serologya | |||
| Nonreactive | 13 | 15 | 28 (3.6) |
| Reactive low titer | 181 | 191 | 372 (48.3) |
| Reactive high titer | 194 | 177 | 371 (48.1) |
| SARS-CoV-2 infection prior to study enrollment | |||
| Yes—≥90 d previously | 149 | 196 | 345 (45) |
| Yes—within the previous 90 d | 50 | 34 | 84 (11) |
| Yes—did not report date of infection | 22 | 15 | 37 (4.8) |
| No | 167 | 138 | 305 (40) |
| Vaccine type among those fully or partially vaccinated at study enrollment | |||
| CoronaVac | 313 | 315 | 628 (80.1) |
| AstraZeneca | 9 | 7 | 16 (2.08) |
| Unvaccinated | 66 | 61 | 127 (17.3) |
| HCW-Weeks | New Cases | Attack Rate (per 1000 HCW-Weeks) | Odds Ratio | 95% CI | |
|---|---|---|---|---|---|
| Sex | |||||
| Male | 676 | 4 | 5.92 | Ref | … |
| Female | 2385 | 12 | 5.03 | 0.85 | .28–2.63 |
| Age | |||||
| <60 y | 2953 | 15 | 5.08 | Ref | … |
| ≥60 y | 108 | 1 | 9.26 | 1.82 | .24–13.99 |
| Occupation | |||||
| Nurse assistant (nurse technician) | 1302 | 9 | 6.91 | 2.63 | .57–12.21 |
| Registered nurse | 469 | 2 | 4.26 | 1.62 | .23–11.53 |
| Security officers | 273 | 1 | 3.66 | 1.39 | .13–15.38 |
| Environmental cleaning staff | 259 | 2 | 7.72 | 2.94 | .41–20.99 |
| Others | 758 | 2 | 2.64 | Ref | … |
| Number of healthcare facilities employed at | |||||
| 1 | 2348 | 14 | 5.96 | Ref | … |
| 2 | 693 | 2 | 2.89 | 0.48 | .11–2.13 |
| ≥3 | 20 | 0 | 0 | … | … |
| Comorbidities | |||||
| No | 2627 | 11 | 4.19 | Ref | … |
| Yes | 434 | 5 | 11.5 | 2.77 | .96–8.02 |
| Household size | |||||
| <5 | 2649 | 15 | 5.66 | Ref | … |
| ≥5 | 412 | 1 | 2.43 | 0.43 | .6–3.24 |
| Prior SARS-CoV-2 infectiona | |||||
| No reported previous infection | 1202 | 10 | 8.32 | Ref | … |
| Previous infection | 1859 | 6 | 3.22 | 0.39 | .14–1.06 |
| Baseline serologyb | |||||
| Nonreactive | 109 | 1 | 9.17 | Ref | … |
| Reactive low titer | 1473 | 9 | 6.11 | 0.67 | .08–5.29 |
| Reactive high titer | 1479 | 6 | 4.06 | 0.44 | .05–3.67 |
| Vaccination status | |||||
| Unvaccinated | 511 | 3 | 5.87 | Ref | … |
| Partially vaccinated | 247 | 1 | 4.05 | 0.69 | .07–6.65 |
| Fully vaccinated | 2303 | 12 | 5.21 | 0.89 | .25–3.15 |
| Hospital unit | |||||
| Non-ICU | 2340 | 10 | 4.27 | Ref | … |
| ICU | 721 | 6 | 8.32 | 1.95 | .71–5.40 |
| Attack rates for weekly exposures | |||||
| Direct contact with a patient with suspected or confirmed COVID-19 | |||||
| No | 1841 | 12 | 6.52 | Ref | … |
| Yes | 1114 | 4 | 3.56 | 0.55 | .18–1.71 |
| Exposure to a confirmed or suspected COVID-19 case outside of hospital | |||||
| No | 2959 | 14 | 4.73 | Ref | … |
| Yes | 102 | 2 | 19.61 | 4.14 | .94–18.76 |
| Direct contact with patient environment | |||||
| No | 1578 | 9 | 5.7 | Ref | … |
| Yes | 606 | 2 | 3.3 | 0.58 | .12–2.68 |
| Highest procedure risk category that week | |||||
| No AGPsc | 2184 | 12 | 5.49 | Ref | … |
| Low-risk | 42 | 1 | 23.81 | 4.33 | .56–34.8 |
| Medium-risk | 191 | 2 | 10.47 | 1.92 | .43–8.62 |
| High-risk | 644 | 1 | 1.55 | 0.28 | .04–2.17 |
DISCUSSION
In this large prospective cohort of 771 HCWs with high vaccine coverage, we identified 16 incident SARS-CoV-2 infections, 12 of which were among fully vaccinated individuals, and approximately 50% of those having the infection within 2 months after being fully vaccinated with 2 doses. During the 3-week follow-up period, the overall attack rate was 5.1 per 1000 HCW-weeks: 5.2 for fully vaccinated HCWs and 5.9 for unvaccinated HCWs.
Environmental cleaning staff experienced the highest SARS-CoV-2 attack rate (7.7 per 1000 HCW-weeks); although not significant in our study, it mirrors findings from similar studies conducted in Brazil and the United States [38, 39]. This could be attributable to differences in training among these employees, who are often outsourced and have low wages with high turnover rates, which may lead to a lack of training aligned with IPC guidelines. Environmental cleaning staff may also have other exposures for increased transmission risk outside of the hospital because of lower socioeconomic status, such as taking public transportation, working at additional healthcare facilities, and household crowding. In contrast, HCWs who reported direct contact with a patient with suspected or confirmed COVID-19 had a lower risk of infection (OR: .55; 95% CI: .18–1.71) than those who did not, despite being directly involved in patient care. Healthcare workers in direct contact are not necessarily at a higher risk of infection, possibly due to their heightened awareness of transmission risk. Despite the nonsignificant result, training for all individuals working in hospital environments, regardless of direct patient contact, is important to mitigate potential exposures.
We did not find any significant association between COVID-19 and healthcare exposures in our study. There are many possible explanations to our lack of significant findings. The low attack rate in this cohort may limit the statistical power necessary to reveal any significant associations between exposure and risk of infection. The attack rate in our study is notably lower than what was observed in Manaus just a few months prior (76% in the general population in the month of October 2020 vs 2% among our cohort participants) [40], possibly due to the short follow-up period and high immunity from prior infection or vaccination at baseline. At enrollment, 74% were fully vaccinated and 61% self-reported prior SARS-CoV-2 infections. Baseline serology revealed that 96% of HCWs had reactive IgG titer values (>35.2 BAU/mL) to the S1 or spike protein, half of which were high (>260 BAU/mL), which could be due to either infection or vaccination. Additionally, enrollment for this study began approximately 3 months after the Gamma variant was first identified and 2 months after a large wave of SARS-CoV-2 infections that devastated the region (Figure 1). These data suggest that our HCW cohort had some immunity (through prior infection or vaccine) against SARS-CoV-2 infection at enrollment, likely protecting against the Gamma strain.
In this study, prior infection and baseline serology levels were stronger protective factors compared with vaccination, although none were statistically significant. This observation could be the result of several factors: (1) the type of vaccine (inactivated whole-virus vaccine—CoronaVac), which has an efficacy of 51% against symptomatic disease compared with more than 90% for molecular RNA vaccines used in the United States [40], and (2) interaction between serology, prior infection, and vaccination since those with prior infection and subsequent vaccination might exhibit a more robust immune response post-vaccination with high levels of antibodies in our study baseline. Although vaccination was not a statistically significant protective factor against COVID-19 infection, those who were vaccinated had no fever, presumptive low viral load estimated by the high Ct values, and rapid clearance, with the subsequent nasal swab at the 1-week follow-up being negative. These findings are similar to other cohort studies in other countries, which have also shown faster viral clearance time among vaccinated individuals compared with unvaccinated individuals [41, 42].
Our study has many strengths, notably a large cohort of HCWs in a highly affected COVID-19 region. The collection of respiratory samples for 3 weeks regardless of symptoms allowed for the identification of asymptomatic and mildly symptomatic individuals. Despite these strengths, the following limitations should be noted:
Lastly, variables such as prior infection and PPE use were self-reported and have the potential for recall bias. Despite these limitations, the study describes important HCW characteristics, including IPC practices, in the country in Latin America with the second-highest number of COVID-19–related deaths [44].
In conclusion, because HCWs are at high risk of infection during pandemics and epidemics of respiratory pathogens, cohort studies involving these individuals can provide valuable information on infection burden and risks. Protection of HCWs is critical to keep healthcare systems operating and delivering healthcare to the population. Our study corroborates the published literature demonstrating a low COVID-19 attack rate among a highly vaccinated population, with a nonsignificant suggestion of a more rapid viral clearance among those vaccinated. It also shows the need for enhanced IPC training across all cadres of HCWs, particularly among nonclinical staff such as environmental cleaners, coupled with adequate availability of PPE to avoid suboptimal prevention practices during patient care activities at the peak of a pandemic. By considering the lessons learned from this study and prioritizing surveillance, education, and vaccination, we can better equip ourselves to mitigate the impact of future outbreaks and safeguard public health.