Journal Pre-proof Risk of SARS-CoV-2 infection in healthcare workers with inflammatory bowel disease: a case-control study

:


INTRODUCTION
Coronavirus disease 2019 (COVID-19) is a respiratory illness caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) [1]. SARS-CoV-2 has been responsible of a massive outbreak since early 2020 and is still a daily concern for worldwide healthcare systems [2].
The risk of COVID-19 or COVID-19-related mortality in patients with IBD have been widely evaluated in many studies and a recent meta-analysis of 14 studies which included 50,706 patients with IBD [3]. The prevalence of COVID-19 in patients with IBD was low and accounted for 1% of patients through October 2020.Whereas it had been presumed that patients who are immunosuppressed would be at higher risk for COVID-19 and severe COVID-19, only the use of steroids seems to impact the risk of COVID-19 [3,4].
Healthcare workers are exposed to a substantial risk for acquiring COVID-19 due to daily and close contacts with infected patients and asymptomatic carriers of SARS-CoV-2. In a recent systematic review with meta-analysis, a total of 97 studies, reported a prevalence of 11% [7%-15%] in healthcare workers and severe COVID-19 in 5% [3%-8%] [5]. In a retrospective case-control study, we showed that healthcare workers with IBD did not have an increased risk of severe infection compared with other patients with IBD, except for tuberculosis [6].
We thus conducted a multicenter case-control study in a real-life setting aiming to assess the incidence rate of COVID-19 in healthcare workers with IBD compared with other non-healthcare worker patients with IBD, and to identify the predictors of COVID-19.
J o u r n a l P r e -p r o o f

Study population
The present study was a follow-up extension of a retrospective observational multicenter case-control study conducted in 17 French and Belgian academic centers affiliated with the Groupe d'Etude Thérapeutique des Affections Inflammatoires du tube Digestif (GETAID). In this study, 482 patients with IBD who were healthcare workers and 482 controls patients were included from the MICISTA registry, a tertiary monocentric clinical database of all consecutive patients with IBD at Saint-Antoine Hospital (Paris, France) [7,8].
Patients were identified from personal databases and/or a standardized hospital inpatient diagnosis datasets. Occupational status of patients with IBD was collected in IBD databases upstream of this study. From January 2021 to October 2021, investigators were asked to report patients previously included in their center who were still followed up until December 31 st 2020.
The protocol was approved by the Henri Mondor Ethics Comittee / Institutional Review Board (N°0011558-2020-070) and the Commission Nationale de l'Informatique et des Libertés (CNIL N°916056). All authors had access to the study data and reviewed and approved the final manuscript.

Data collection
Previously collected data were retrieved and updated until December 31 st 2020. The recorded data included a detailed account of the IBD diagnosis and history, smoking status, IBD phenotype according to the Montreal classification, medical and surgical treatment history and any serious infection history. For each of the patients, year 2020 was divided into semesters. For each semester, which was independently analyzed, the occurrence of COVID-19 and severe COVID-19, smoking status, physician global assessment of IBD activity (active J o u r n a l P r e -p r o o f or not), weight and immunosuppressive therapy (e.g., steroids, aminosalicylates, thiopurines, methotrexate, anti-TNF therapy, ustekinumab, anti-integrin therapy and tofacitinib) were assessed [9]. A semester was considered as a treatment semester if the patient received steroids, immunomodulator and/or anti-integrine therapy during at least 3 months within the studied semester.

Case-control study
In the previous study, controls were selected randomly within the MICISTA registry to match to the healthcare worker cases (1:1). MICISTA is an electronic database of the gastroenterology department of Saint-Antoine Hospital. All patients seen in the institution from 1994 are included in the database. Data regarding medical and IBD history and followup are prospectively coded in the system. Case-control matching was based on gender, birth year (± 2.5 years), type of IBD and IBD diagnosis calendar (± 2.5 years). After exclusion of 156 cases and 143 controls (Figure 1), 221 case-control couples were still available. Onehundred and five cases were rematched with 105 controls without redundancy. In total, 326 cases and 326 controls (one control for one case) were included in the present study.

Outcomes
The main outcome measure was to compare the overall incidence of COVID-19, defined as symptomatic SARS-CoV-2 infection with a positive reverse transcription polymerase chain reaction (RT-PCR) test for SARS-CoV-2 on nasopharyngeal swab and/or a positive serological test for SARS-CoV-2 without prior vaccination. Secondary outcome was the overall incidence of severe COVID-19, defined as any COVID-19 requiring hospitalization, intensive care unit stay or death, and the impact of patients' characteristics, occupational status and ongoing treatment on the risk of COVID-19 and severe COVID-19. The rates of overall COVID-19 and severe COVID-19 were expressed for 100 patient-semester.

J o u r n a l P r e -p r o o f
Continuous data were expressed as means ± standard deviations or medians (interquartile range) whereas nominal and ordinal data were expressed as numbers and percentages. Continuous data were compared using the Chi-square test or the Fischer's exact test whenever appropriate. Parametric data were compared using the Mann-Whitney tests and Wilcoxon's matched-pair signed-rank test as appropriate. COVID-19 -free survival was calculated using the Kaplan-Meier method. To determine risk factors of COVID-19, a multivariate analyses using binary logistic regression models were performed and adjusted according to the results of univariate analysis with an ascending stepwise procedure using the Wald test. Quantitative values were converted to qualitative values using the dichotomy from the median value in two distinct groups of equal size. Variables with p < 0.10 in the univariate analysis were considered to be potential adjustment variables for the multivariate analysis. All analyses were two-tailed, and p values less than 0.05 were considered significant. Statistical evaluations were performed using SPSS statistical software (SPSS Inc., v23, Chicago, IL, USA).
The control group included 326 non-healthcare worker patients with IBD. Patient demographic data, baseline diseases characteristics and medication history of the control group are also listed in Table 1. The two groups were well balanced except for a slight difference in the age at diagnosis distribution according to the Vienna classification but without significant difference on the mean age at diagnosis (25.9 ± 10.2 vs. 25.5 ± 10.4, p = 0.21). Healthcare workers were less likely to smoke (9.3% vs. 14.4%, p = 0.05) and were also less frequently confined at home during the lockdown periods (16.6% vs. 14.4%, p < 0.001) which is consistent with their occupational status.
In total, we collected six severe COVID-19 in four case and two control patients accounting for an incidence rate of 0.5 ± 6.8 overall severe COVID-19 per 100 patient-J o u r n a l P r e -p r o o f semesters ( Table 2). No deaths were noted in healthcare workers and the control group. One patient required admission in intensive care unit and subsequently recovered. No difference was noted between the healthcare workers and the control group regarding the incidence rate of severe COVID-19 (0.6 ± 7.8 vs. 0.3 ± 5.5 per 100 patient-semesters, p = 0.42) ( Table 2).

Predictors of COVID-19
Predictors of overall severe infection were assessed in the entire study population including cases and controls. In univariate analysis, patients with BMI > 25 kg/m² (p = 0.05) and > 30 kg/m² (p = 0.03), nurses' aides (p = 0.02), Crohn's disease of the upper GI tract (p = 0.003) and with proctitis E1 (p = 0.06) ( J o u r n a l P r e -p r o o f DISCUSSION In the present study, we investigated for the first time the risk of healthcare workers with IBD to develop COVID-19 compared to control patients with IBD. In our cohort, the prevalence of COVID-19 was 8.7% of patients with IBD during year 2020. No difference was observed according to occupational status. We did not found any impact of IBD treatment on the risk of COVID-19. Lastly, we observed that patients with IBD were more likely to develop COVID-19 when their BMI was > 30 kg/m². Exposure to pathogen is a risk factor for opportunistic infection in the immunocompromised population [10]. Healthcare workers have higher susceptibility to respiratory infections in general and more specifically, zoonotic coronavirus outbreaks such as SARS, MERS and COVID-19 [11,12]. In a recent systematic review of 97 studies (all published in 2020) concerning 230,398 healthcare workers, the estimated prevalence of COVID-19 was 11% [7][8][9][10][11][12][13][14][15] using RT-PCR tests and 7% [4][5][6][7][8][9][10][11] using serological tests while severe COVID-19 occurred in 5% [3][4][5][6][7][8] of healthcare workers [5]. Being at the frontline response to COVID-19 results thus at higher risk of acquiring the disease. As an example, 2,600,498 patients were diagnosed with COVID-19 in France through 31 December 2020, accounting for 3.9% of the whole French general population [13]. In our study, the prevalence of COVID-19 was 9.8% which is similar to the latter data of patients without IBD. The distribution of COVID-19 was homogeneous across different healthcare workers' categories.
Surprisingly, there was no difference between healthcare workers with IBD (9.8%) and controls with IBD (7.7%, p = 0.41). France has been hit early during the European SARS-CoV-2 outbreak, early set-up of protective measure during year 2020 may explain such finding [14]. In the present study, the only healthcare workers' category who were at higher risk of COVID-19 was nurses' aides (14.0% vs. 5.5%, p = 0.02). We speculate that the closest contact with in-hospital patients that is associated with the work of nurses' aides may explain J o u r n a l P r e -p r o o f this difference. Independently of occupational status, we found an increased risk for our control population with IBD. In a recent systematic reviews with meta-analysis including 23 studies and 51,463 patients with IBD, through October, the prevalence of SARS-CoV-2 infection was 1.01% [0.92-1.10] [15]. It is conceivable that such results may reveal heterogeneity in the individual risk of getting COVID-19 for patients with IBD according to epidemiological differences of the COVID-19 outbreak and the anti-COVID-19 measures across countries and differences in the access to microbiological diagnosis using RT-PCR and/or serology. Notably, there was no difference between groups considering rates of intravenous biologics (infliximab and vedolizumab) administered at hospital infusion centre (74 (22.7%) in both groups, p = 1.00).
In the recent meta-analysis on the risk of COVID-19 in patients with IBD, 9 out of 23 studies reported specific outcomes according to ongoing treatments. Worse outcomes were noted in patients treated with steroids and aminosalicylates and better outcomes for those treated with biologics and immunosuppressants [15]. This could may be result from a potential effect of biologics and immunosuppressant on SARS-CoV-2 infection which has been demonstrated with tofacitinib and tocilizumab [16,17]. We did not observe such differences in our cohort. However, we have the relatively low number of patients did not allow us to performed subgroup analysis according to various IBD treatments.
In the present study, the only predictor of COVID-19 was obesity, defined as a BMI > 30 kg/m². Obesity has been repeatedly reported as a risk factor of COVID-19 and severe COVID-19 in various cohort and epidemiological studies [18][19][20][21][22]. The link between obesity and worse COVID-19 outcomes is complex including dysregulated immune response and altered mechanics of lungs and chest wall. Obesity is more and more reported among patients with IBD, accounting for approximately 15 to 40% of adults with IBD in cross-sectional studies [23] , [24]. Those considerations are not specific of SARS-CoV-2 but has also been J o u r n a l P r e -p r o o f previously pointed out for influenza virus and other respiratory viruses [25]. However, SARS-CoV-2 exerts an even higher cytokine storm in patients with underlying metabolic syndrome, diabetes and obesity [21].
Whereas, we included a large number of healthcare workers with IBD, we acknowledge a number of limitations in this study. First, the size of the present study may be too small to assess statistical difference between both groups. Second, data collection was retrospective during year 2021. However, all the participating centres are tertiary care centre with standardized clinical, biological, endoscopic and morphological prospective follow-up that lower the impact of such bias. Third, we focused on COVID-19 defined as symptomatic SARS-CoV-2 infection with positive RT-PCR test for SARS-CoV-2 on nasopharyngeal swab and/or a positive serological test for SARS-CoV-2 without prior vaccination. Nonetheless, this bias should be balanced in both groups. Fourth, we lost one third of the original cohort in both the healthcare worker and control groups and had to subsequently proceed to a rematch process of cases and controls. However, both groups were ultimately well balanced with few significant differences.
We concluded that healthcare workers with IBD did not exhibit an increased risk of COVID-19 compared with controls. Special attention should be given to nurses' aides with regard to closest contact with patients and underlying risk of transmission of SARS-CoV-2.
Those data are reassuring as well as the low incidence of severe COVID-19 either in healthcare workers and control patients with IBD.     J o u r n a l P r e -p r o o f J o u r n a l P r e -p r o o f J o u r n a l P r e -p r o o f