Infectiousness of SARS-CoV-2 breakthrough infections and reinfections during the Omicron wave

Using detailed epidemiologic data from SARS-CoV-2 surveillance within the California state prison system, we found that vaccination and prior infection reduced the infectiousness of SARS-CoV-2 infections during an Omicron-predominant (subvariants BA.1 and BA.2) period. Vaccination and prior infection were each associated with similar reductions in infectiousness during SARS-CoV-2 infection, and, notably, additional doses of vaccination (for example, booster doses) against SARS-CoV-2 and more recent vaccination led to greater reductions in infectiousness. Of note, reductions in transmission risk associated with vaccination and prior infection were found to be additive, indicating an increased benefit conferred by vaccination for reducing cases’ infectiousness even after prior infection. Irrespective of vaccination and/or prior natural infection, SARS-CoV-2 breakthrough infections and reinfections remained highly infectious and were responsible for 80% of transmission observed in the study population, which has high levels of both prior infection and vaccination. This observation underscores that vaccination and prevalent naturally acquired immunity alone will not eliminate risk of SARS-CoV-2 infection, especially in higher-risk settings, such as prisons.

Prior studies during the Delta variant wave and before widespread booster vaccination are mixed on whether SARS-CoV-2 breakthrough infections in vaccinated individuals are potentially less infectious^6,7,8 or equally infectious^9,10 to primary infections. In more recent household contact studies during the Omicron variant era^11,12,13, vaccination often led to reduced SARS-CoV-2 infectiousness. Several factors may have enhanced our ability to observe statistically meaningful findings in the present study. The risk of transmission among close contacts in the prison setting and consistency in contact structure, especially in light of increased transmissibility of the Omicron SARS-CoV-2 variant, may have enhanced statistical power in our sample. Relatedly, a higher proportion of index cases in our sample were previously vaccinated or infected, further enhancing the opportunity to compare transmission risk from vaccinated or unvaccinated index cases and from those who were previously infected or previously uninfected.

A key result is that the vaccine-mediated reduction in infectiousness of SARS-CoV-2 breakthrough infections appears to be dose dependent. Each dose of the vaccine provided an additional average 11% relative reduction in infectiousness, which was mostly driven by residents with a booster dose. The findings of this study support the indirect effects of COVID-19 vaccination (especially booster doses) to slow transmission of SARS-CoV-2 and build on evidence of the direct effects of COVID-19 vaccination²³ to emphasize the overall importance of COVID-19 vaccination. The public health implication of these findings is further support for existing policy using booster doses of vaccination²⁴ to achieve the goal of lowering population-level transmission. The impact of additional bivalent vaccine doses, which are now authorized for individuals over 5–6 years of age²⁵, on transmission should be a priority for further study. Additional considerations about the timeliness of vaccine doses are also necessary, as we found that index cases with more distant history of COVID-19 vaccination had a higher risk of transmission of infection to close contacts. Given this finding, this study raises the possibility of timed mass vaccination in incarcerated settings during surges to slow transmission.

The findings from this study have direct implications in addressing COVID-19 inequities in the incarcerated population through additional vaccination. In California state prisons at the time of this study, although 81% of residents and 73% of staff have completed a primary vaccination series, only 59% of residents and 41% of staff have received the number of vaccination doses recommended by the Centers for Disease Control and Prevention based on their age and comorbid medical conditions²⁶. Our findings also provide a basis for additional considerations for housing situations of cases based on prior vaccination and infection history in future surges and can be used alongside other measures, such as depopulation and ventilation interventions, to protect incarcerated populations.

However, this study also underscores the persisting vulnerability to COVID-19 among residents and staff in correctional settings despite widespread vaccination, natural immunity and use of non-pharmaceutical interventions. The overall attack rate of SARS-CoV-2 in the study population (who were generally moved into isolation after symptoms or a positive test) was 30%, and index cases with breakthrough infections or reinfections remained highly infectious, which call into question the ability of high vaccination rates alone to prevent all SARS-CoV-2 transmission in correctional settings. In the United States, which incarcerates more residents per capita than any other country in the world²⁶ and has a quarter of the world’s incarcerated population, correctional settings are characterized by poorly ventilated facilities, populations with increased rates of comorbid health conditions, high-risk dormitory housing and overcrowding^18,27,28,29. Given the inability of current efforts to reduce transmission of SARS-CoV-2, decarceration efforts may be the most likely to have substantial effects on reducing cases.

The secondary attack rate in this study was on the lower end of published estimates when comparing to household studies. Of note, the secondary attack rate of the SARS-CoV-2 Omicron variant in recent household studies ranges from 29% to 53%^11,12,13, in contrast to a 30% attack rate in this study. The prison environment has distinct epidemiologic differences to households. The dense living environment increases the likelihood of transmission in the prison environment compared to a household, whereas the frequent asymptomatic testing (with isolation of positive cases) in the prisons likely reduced the exposure time and subsequent transmission risk compared to households. The transmission of the prison cell is also likely more uniform than a household.

Strengths of this study include access to detailed records of all residents in the California state prison system, encompassing individuals’ prior COVID-19 vaccine receipt and prior natural infection history (based on frequent testing throughout the pandemic), as well as a social network given record of where residents slept each night over the study period. We use a consistent definition of social contact between the index case of COVID-19 and close contact based on the uniformity of cell type. The frequent testing ensures early identification of infections and systematic capture of asymptomatic and symptomatic infections to avoid bias by participants’ immune status (which could affect temporal onset of symptoms). The risk of misclassification of close contacts is low given that most follow-up testing in close contacts occurred well after first exposure to an index case (Supplementary Notes). The large sample size facilitates analyses of the contribution of combinations of prior vaccination statuses and natural infection on risk of transmission, including analyses examining the impact of booster doses.

Limitations should also be considered. We cannot exclude the possibility of some residual confounding (for example, behavioral differences that affect risk of transmission) between individuals who were vaccinated against SARS-CoV-2 and those who were unvaccinated. There is a possibility that close contacts who test positive for SARS-CoV-2 were not infected by their assigned index case but, instead, by interaction with infectious individuals outside of their cell. However, this misattribution would be expected to dampen apparent associations of transmission risk with index cases’ vaccination status and infection history but not bias the relative estimates. To further address the risk of misattribution, we adjusted for background SARS-CoV-2 incidence and matched contact pairs by facility and time. Our study population is a subset of the entire incarcerated population in California and may not represent all incarcerated settings. Studies of SARS-CoV-2 infectiousness may be subject to biases^30,31,32. The strict inclusion and exclusion criteria in this study may introduce bias into the analysis, although we performed sensitivity analyses on these criteria with overall consistent findings. We also adjusted for prior infection in analyses to account for potential concerns about differential susceptibility related to prior infection in vaccinated versus unvaccinated individuals. Given limited SARS-CoV-2 testing capacity early in the pandemic and some residents’ decision to decline testing, it is possible that infections among some residents may not have been captured, although such misclassification would be expected to bias our findings to the null. SARS-CoV-2 testing was variable over time in the prison system, with periods of routine weekly testing and other periods of reactive testing; however, periods without reactive testing align with times during which SARS-CoV-2 was unlikely to be circulating at high levels within the facilities, suggesting that this is unlikely to bias results substantially. The study findings on boosters may also be related to recent vaccination effects. This study design did not provide a basis for identifying effects of vaccination and prior infection on risk of acquiring SARS-CoV-2 among close contacts, although we did adjust for prior infection and vaccination in close contacts in the primary analysis. Of note, vaccine effectiveness against infection among incarcerated persons has been reported within this population during earlier periods^33,34. We do not have a detailed record of person-level masking, symptoms, cycle thresholds for polymerase chain reaction (PCR) testing or serologic testing. During the study, the predominant Omicron subvariants in California and California prisons were BA.1 and BA.2 based on genomic surveillance, although we did not genotype every SARS-CoV-2 isolate in this study.

This study demonstrates that breakthrough COVID-19 infections with the Omicron variant remain highly infectious but that both vaccination and natural infection confer reductions in transmission, with benefit of additional vaccine doses. As SARS-CoV-2 breakthrough infections and reinfections become the predominant COVID-19 case, this study supports the importance of booster doses in reducing population-level transmission with consideration of mass timed vaccination during surges, with particular relevance in vulnerable, high-density congregate settings.