A number of underlying medical conditions such as cardiovascular disease, diabetes, and obesity increase the risk for COVID-19 infection and severity.^1,2 Sleep conditions also are important risk factors associated with increased prevalence and worse outcomes for COVID-19 disease; emerging evidence demonstrates that obstructive sleep apnea^3-6 and more recently, insomnia and reduced sleep duration are linked to an increase in prevalence and worse COVID-19 disease outcomes.⁷ Circadian rhythmicity is important in the regulation of immune system and inflammatory processes.^8,9 Chronotype is a trait that strongly reflects circadian timing as well as social and homeostatic factors.¹⁰ Recently, evening chronotype has been associated with greater risk of poor health with increased rates of obesity, cardiovascular disease, and diabetes.^11,12 However, there have been few studies of whether a specific diurnal preference or chronotype is associated with a greater risk for COVID-19 disease. In a single center study in France, social jet lag as a marker of evening chronotype was associated with a higher rate of COVID-19.¹³ In contrast, the COVID-19 Outbreak Public Evaluation (COPE) Initiative reported that morning in comparison to evening chronotype was associated with a higher rate of COVID infection;⁷ while in a study using UK Biobank participants, no association was observed between a specific chronotype and COVID-19.¹⁴ Thus, the impact of chronotype on COVID-19 has not been determined.

During the height of the COVID-19 pandemic, sleep patterns changed. As a result of lockdowns and social restrictions, stress, meal timing, anxiety and depression, social isolation, and the migration to remote working, both increases and decreases in sleep duration as well as changes in sleep timing occurred.^15-17 The presence of these factors may allow the development of circadian misalignment which increases risk of cardiovascular and metabolic disorders, and importantly infections. Circadian misalignment is a characteristic of shift workers who have been shown to have an increased risk of developing respiratory infections.^18,19 It has been suggested that shift workers are at greater risk for COVID-19.^20-22 This has been confirmed by several studies observing that night shift work is associated with higher rates of COVID-19 infection and hospitalization.^23-27

Investigations examining whether the adverse effect of shift work on health is modified by chronotype suggest that shift workers with evening chronotype may have greater risk for diabetes and obesity.¹² However, there have been no assessments of whether risk of infection and in particular COVID-19 is similarly affected.

In this study, we examined whether both morning and evening chronotypes were more likely to have had a COVID-19 infection in comparison to those who were intermediate chronotypes. Additionally, we examined whether there is a synergistic association between chronotype and shift work, with COVID-19 infection. To accomplish this, we used data from the first four 2022 waves of COPE Initiative (http://www.thecopeinitiative.org/), a program focused on accumulating data on public attitudes, behaviors and beliefs related to the COVID-19 pandemic from large scale, demographically representative samples.

Table 1 shows the bivariate associations between COVID-19 infection status, diurnal preference, and co-morbid medical, demographic, and social characteristics of the cohort. As shown in the Supplementary Digital Content, of the 19821 participants, 7932 (40.0%) had at least one COVID-19 infection. Morning and evening diurnal preferences were observed in 31.9% and 19.4% of participants respectively. Among COVID-19 positive in comparison to COVID-19 negative participants, there was a higher percentage of morning types (32.6 vs 31.4%) and a lower percentage of evening types (17.2 vs. 20.9%, p≤0.001). Participants who were retired or unemployed were less likely to have had a COVID-19 infection (12.5 vs. 30.8%, p≤0.001). COVID-19 positive in comparison to COVID-19 negative participants were younger (40.2 ± 15.8 vs. 50.4 ± 18.0 y, p≤0.001), had a slightly shorter self-reported sleep duration (6.8 ± 1.6 vs. 6.9 ± 1.7 h, p≤0.001), had a slightly larger percent of time working remotely (36 ± 34 vs. 32 ± 38%, p≤0.001), were slightly more overweight (BMI: 28.7 vs. 28.3 kg/m², p<0.001) had more comorbidities (1.7 ± 2.4 vs. 0.8 ± 1.2, p≤0.001), and were less likely to have received a COVID-19 booster vaccination (27.6 vs. 46.4%, p≤0.001). COVID-19 positive participants also were more likely to be Hispanic, not retired and have an income more than $50,000 per year.

The unadjusted associations between COVID-19 infection and shiftwork status, and employment sector are shown in Table 2. In comparison to the Healthcare sector, employees working in Construction, Financial and Protective Services had higher rates of infection while those working in Factory/Manufacturing, Retail, Transportation and Other sectors had lower rates. In comparison to Office/Administration employees, Arts/Entertainment/Media, Cleaning/Maintenance, Factory/Manufacturing, Food Industry, Healthcare, Protective Services, Retail and Transportation had a lower prevalence rate of Day Shift only positions. In contrast, those working in Other sectors had higher rates of Day Shift only positions.

Table 3 displays the logistic regression models for the association between having had a COVID-19 infection and either a morning or evening diurnal preference in comparison to a reference group without a definite preference. In models including demographics, comorbidities, vaccination status, and socioeconomic factors, morning preference was related to a 20% greater likelihood of COVID-19 infection. In contrast, evening preference was 20% less likely to be linked to a COVID-19 infection. These associations were not materially affected by the type of work shift, percent time working remotely or after adjustment for sleep duration.

In Table 4 are the bivariate associations between COVID-19 infection and diurnal preference, and co-morbid medical, demographic, and social characteristics of the cohort limited only to members who were employed (N=11114, Figure). In comparison to the entire cohort, participants were slightly younger, more likely to be male, and had a higher income. Participants who worked a mixture of shift types had a higher prevalence rate of infection in comparison to those who worked only day shifts or only night shifts (p<0.001).

Table 5 shows the fully adjusted logistic regression model which adjusted for percent remote work and sleep duration documenting the association between COVID-19 infection and diurnal preference, work shift category and the interactions between diurnal preference and work shift category. In comparison to neither definite morning nor evening preference, there was a 18% greater likelihood of infection in those with morning preference and a 16% lower likelihood in those with an evening preference. Both were comparable to the aORs observed for the overall cohort. Addition of employment sector did not substantially alter these findings. In comparison to working only a day shift, a greater risk of COVID-19 infection was observed in those working a mixture of day and evening shifts (aOR: 1.25, 95% CI: 1.10-1.41); working a mixture of day, evening, and night shifts approached significance (aOR: 1.18, 95% CI: 0.97-1.44). Working only night shifts was not associated with a higher prevalence of COVID-19 infection. The overall interaction between diurnal preference and work shift category was significant (Wald χ²=13.864, p=0.031). Individual interaction terms indicated that there was a higher likelihood of infection in participants who had a morning preference and working a combination of day and evening shifts (aOR: 1.87, 95% CI: 1.28-2.74); the association with those with a morning preference and working a mixture of day, evening, and night shifts approached significance (aOR: 1.72, 95% CI:0.98-3.02).

Sensitivity analyses indicated that our findings were similar using slightly stricter or more liberal definitions of COVID-19. However, a definition that required a positive test only lost statistical significance.

In this study, morning and evening chronotypes were found to be differentially associated with COVID-19 infection; morning types were more likely and evening types were significantly less likely to be infected. Additionally, in comparison to working only day shifts, working a combination of different shift types, as opposed to day shifts only was associated with a higher prevalence of COVID-19 infection. This finding of higher infection rate was primarily observed in individuals with a morning chronotype who worked a combination of different shifts.

Morning chronotypes were more susceptible to COVID-19 infection than evening chronotypes; this finding is consistent with earlier preliminary results from the same cohort.⁷ This observation stands in contrast to a large body of evidence demonstrating that the prevalence of cardiovascular disease risk factors and metabolic disorders is higher in evening chronotypes.^11,12 Furthermore, our results are distinct from the limited studies that also have assessed the association between chronotype and COVID-19. In a single center study, social jet lag as a marker of evening chronotype was found to be associated with a higher rate of COVID-19.¹³ In contrast, no association between chronotype and COVID-19 infection was observed in two separate analyses of the UK Biobank^14,26 as well as in an analysis of long COVID from the Nurses’ Health Study.³² Additionally, to our knowledge, there have been no previous studies of an association between chronotype and infection risk from other pathogens.

Despite the paucity of clinical investigations related to chronotype and infection risk, there are mechanisms that could explain our finding of an association between morning chronotype and COVID-19 infection. A reduction in sleep duration has been linked to increased susceptibility to infection.¹⁹ However, our observation remained robust after controlling for sleep duration in our modelling. Alternatively, various components of the immune system exhibit circadian rhythmicity.³³ Secretion of inflammatory cytokines, trafficking myeloid and lymphocyte subsets and maturation of leukocytes exhibit a circadian rhythm with the aggregate response favoring a proinflammatory response while asleep.³³ Some,^34,35 but not all^36,37 have found that time of day influences the antibody response to COVID-19 vaccination. Therefore, it is possible that immunity to COVID-19 is less robust in morning chronotypes resulting in greater susceptibility to infection.

In this study, the association between working night shift and COVID-19 infection was the same as working day shift. However, working a mixture of shifts, particularly ones that included evening shifts was associated with a higher likelihood of COVID-19 infection. These findings stand in contrast to some previous observations that have noted a positive association between rotating and consistent/permanent night shift work and COVID-19 infection.^24,25,27 However, not all studies have found a significant relationship among those who only worked consistent/permanent night shift.^23,26 Working night shifts and early morning shifts are associated with circadian misalignment as well as shorter sleep duration. Both may increase the propensity for COVID-19 infection as well as other respiratory contagions. However, night shift workers tend to have less interpersonal contact than day shift workers.³⁸ This may mitigate the adverse impact of circadian misalignment on increased risk of COVID-19 infection. Our results suggest that the impact of circadian misalignment and sleep loss, and consequent predisposition towards being infected with COVID-19 is greater among persons who irregularly perform shift work rather than in those who work permanent/consistent day or night shifts.

We observed a significant interaction between chronotype and shift work in which morning diurnal preference together with working a mixture of shifts was associated with the highest aOR of having had a COVID-19 infection. This implies that irregularity in sleep schedule was a major factor in susceptibility to COVID-19 infection in those with morning diurnal preference. Irregularity in sleep-wake patterns has been associated with higher levels of cardiovascular disease biomarkers³⁹ as well as a greater risk of cardiovascular disease.⁴⁰ Using UK Biobank data, sleep irregularity also has been demonstrated to increase risk of COVID-19 infection and its severity.⁴¹ Potential explanatory mechanisms include the impact of sleep irregularity on reducing sleep duration as well as consequences of higher levels of circadian misalignment. Both of these factors can contribute to greater amounts of both acute and chronic inflammation resulting in higher risk of COVID-19 infection.

Several other observations from our study are noteworthy. Remote workers in our study had a slightly higher COVID-19 infection rate which is counterintuitive to the concept of less interpersonal contact should reduce infection risk.⁴² However, this was not confirmed in a recent study where remote work was found to have a higher risk of COVID-19.⁴³ Possible explanations were that remote work led to a false sense of security and less use of personal protective measures such as mask wearing and handwashing as well as the opportunity to engage in other high risk activities such as dining out with others.⁴³ We also noted that higher income was associated with greater COVID-19 infection risk. Although many studies have identified lower income as a risk factor for COVID-19 infection, others have noted the converse or no impact.⁴⁴ Income is associated with many factors including type and place of employment, living conditions, educational attainment, race and/or ethnicity. In our study, these factors favored a slightly greater COVID-19 rate in those with a higher income. Additionally, we observed that there was substantial variation in COVID-19 infection and shiftwork status among various employment sectors. Because of the heterogeneity between and within employment sectors and shiftwork status with respect to interpersonal contact, these differences were expected. We do not believe any of these factors altered our findings related to circadian preference in that they were included in our fully adjusted models.

Our study should be interpreted in the context of several limitations. First, all of the data were self-reported including ascertainment of both diurnal preference (using a single self-report item) and COVID-19 infection. However, sensitivity analyses using different definitions of COVID-19 infection were qualitatively similar to the findings reported herein. In addition, broadly similar to a previous report, we found that morning chronotypes were more prevalent than evening chronotypes, suggesting that our ascertainment of chronotype was acceptable.⁴⁵ Second, our analyses were cross-sectional and causal inference cannot be assumed. Third, although we attempted to adjust for a number of factors known to increase risk of COVID-19 infection, residual confounding is possible.

In conclusion, morning diurnal preference is associated with an increase and evening diurnal preference is associated with a decrease in COVID-19 infection. Neither working day nor night shifts were linked to an increase in prevalence of COVID-19 infection. However, COVID-19 infection was more likely in those working a mixture of shifts, with the greatest risk conferred on those with a morning chronotype. These findings may be informative for developing measures for greater COVID-19 surveillance of shift workers with a morning chronotype, particularly as first responders in healthcare and other settings are often engaged in shift work.

Data are available upon appropriate request to the corresponding author.