Outcomes of COVID-19 in Patients with Inflammatory Bowel Disease: Comparison with Household Members and the Role of IBD Medications

Ali Reza Sima, MD, MHPE1; Bahar Saberzadeh-Ardestani, MD, MPH, MHPE1; Homayoon Vahedi, MD1; Hafez Fakheri, MD2; Fariborz Mansour-Ghanaei, MD3; Iradj Maleki, MD2; Siavosh Nasseri-Moghaddam, MD1; Hasan Vosoghinia, MD4; Mohammad Reza Ghadir, MD5; Ahmad Hormati, MD5,6; Amir Kasaeian, PhD1,7; Amir Reza Radmard, MD8; Bardia Khosravi, MD, MPH, MHPE1; Masoud Malekzadeh, MD1; Sudabeh Alatab, MD, PhD1; Anahita Sadeghi, MD1; Nayyereh Aminisani, PhD9; Hossein Poustchi, MD, PhD1; Elnaz Gonoudi, DDS1; Amir Anushiravani, MD1; Maryam Rayatpisheh, MD1; Jean-Frederic Colombel, MD10; Ryan C. Ungaro, MD, MS10*; Reza Malekzadeh, MD1*


Introduction
Inflammatory bowel diseases (IBDs), comprising Crohn's disease (CD) and ulcerative colitis (UC), lead to chronic inflammation of the gastrointestinal (GI) tract. 1 It is estimated that IBD affects up to 6.8 million of the world population. 1 Immune system dysregulation plays a key role in the disease pathogenesis. 2 Therefore, immunosuppressive therapies such as azathioprine and corticosteroids are frequently used treatments for IBD patients, 3 making them prone to infections. [4][5][6] Case-fatality rates ranging from 2.3%-7.2% have been reported for coronavirus disease 2019 (COVID-19) since its emergence. 7,8 Older age, male sex, and multiple comorbidities have been associated with severe outcomes. [7][8][9] Patients with IBD and their physicians have had concerns about IBD routine care and the effect of IBD and related treatments on COVID-19 susceptibility and outcomes. 10,11 IBD symptoms and COVID-19 may overlap since COVID-19 infection can manifest with gastrointestinal symptoms such as diarrhea. 12,13 A number of studies have found that the COVID-19 incidence rate seems to be equal to or lower among IBD patients. 14,15 Age, uncontrolled IBD, and the number of comorbidities have been associated with adverse outcomes in patients with IBD and COVID-19. 16,17 Viral infections, such as Cytomegalovirus and Herpes Simplex Virus, can lead to more severe complications in patients with IBD taking immunosuppressive drugs. 18 Less is known about the impact of immunosuppressive drugs on COVID-19 in IBD patients.
While immunosuppressive therapy could facilitate the activity of opportunistic organisms in the host, 19 it may reduce severe outcomes in COVID-19 patients by inhibition of cytokine storm. 20,21 Although some studies have suggested a positive association between IBD medications and severe COVID-19 outcomes, others have not seen such an association. 16,17 Studies have shown that immunosuppressive therapies are not associated with adverse outcomes of previous coronaviruses. 22 For better management of patients with IBD during the COVID-19 outbreak, a deeper understanding of the relationship between COVID-19 outcomes and IBD treatments is necessary. Most of the previous literature did not compare IBD patients to patients with similar exposures to COVID-19, such as their household members. This study aims to describe the characteristics of COVID-19 among IBD patients as compared to non-IBD household members and investigate the association between COVID-19 outcome and IBD treatments.

Study Design and Participants
In a multicenter prospective study from six provinces in Iran, recruited from February until April 2020, all IBD patients 18 years and older were enrolled in this study. The six included provinces represent the highest IBD prevalence registered in the Iranian Registry of Crohn's and Colitis (IRCC) and these data are a subset of the IRCC. 23,24 All patients were contacted by phone, and the following data were obtained: medical history and COVID-19 status among patients with IBD. They were also asked about all their household members with regard to COVID-19 and their health status. Age, sex, comorbidities, cigarette or waterpipe smoking, symptoms and outcomes of COVID-19 were collected from non-IBD household members. We used household members of patients with IBD to compare COVID-19 patients with and without IBD since their exposures were relatively similar. All IBD patients, calling the national hotline specified for IBD patients for COVID-19 during this period, were also enrolled. Figure 1 shows the study flow chart.
COVID-19 status was classified as suspected or confirmed. Patients with a history of contact with an infected person and a minimum of three signs/symptoms (fever, cough, dyspnea, dysosmia, and dysgeusia, or lung CT findings consistent with COVID-19) were considered as a suspected case. 25,26 Data from patients with suspected COVID-19 were reviewed by an expert team of radiologists and gastroenterologists and based on that, the COVID-19 diagnosis was confirmed for each COVID-19 case. In addition to the above mentioned COVID-19 cases, patients with positive SARS-CoV-2 nasopharyngeal RT-PCR were considered as confirmed COVID-19 cases.
Demographic characteristics and past medical history of enrolled patients were extracted from the IRCC database. We assessed the following for COVID-19 cases: hospitalization, intensive care unit (ICU) admission, respiratory support, and death at four weeks by telephone calls and medical record review.
The primary outcome was hospitalization due to COIVD-19. Secondary outcomes included ICU admission, intubation and death. Severe COVID-19 was defined as any case of ICU admission, intubation, or death.

Statistical Analysis
Continuous variables are described as mean (standard deviation, SD) or median (interquartile range, IQR), for the variables without normal distribution. Categorical variables are represented as frequencies. One-way analysis of variance (ANOVA test), Kruskal-Wallis test, Bonferroni multiple-comparison test, and Mann-Whitney test were used for comparing means. Pearson chi-square and Fisher's exact test (if needed) were used to compare categorical variables. Crude data are presented for the overall study population with multiple outcomes stratified by the patients' characteristics. Univariable logistic analysis was performed for variables with a P value < 0.05 in the tests for comparing means of continuous variables and distribution of categorical variables. Variables with a P value < 0.05 in univariable logistic analysis that were congruent with prior evidence were included in a multivariable logistic regression model to investigate the independent effect of variables on hospitalization.
Also, we analyzed the association between the outcome of hospitalization and patients' characteristics, including IBD medications, among patients with IBD using multivariable logistic regression.
We performed all statistical analyses using Stata 11.2 for Windows (StataCorp. 2011. Stata Statistical Software: Release 12. College Station, TX: StataCorp LP). For followup pairwise comparisons with the Man-Whitney test, a P value < 0.016 was considered significant; otherwise, a P value < 0.05 was considered significant. Table 1 details the patients' demographic and clinical characteristics. We collected 133 patients with COVID-19, which consisted of 84 IBD patients and 49 household controls. Patients who contacted the hotline comprised 54.7% of IBD patients with COVID-19 There were no significant age and sex differences between patients with and without IBD (P = 0.98 and 0.89, respectively). In the overall population, hypertension and diabetes were the most prevalent comorbidities, and the most commonly reported symptoms were dyspnea, cough, and fever. Dyspnea and cough were significantly more common among patients with IBD. Sixty percent of all IBD patients (38 patients out of 63) and 30% of household controls (13 patients out of 43) developed new GI symptoms (P = 0.002). Also, 1.2% of IBD patients with COVID-19 had household members with COVID-19. Totally, 71.4% of patients with IBD had UC, and more than 50% of patients with IBD were in remission. The most commonly used medication was 5-aminosalicylic acid (5-ASA) (70.24%). Other medications are summarized in Table 1.

Results
The distribution of different outcomes by demographic and clinical data in the overall study population (COVID-19 patients with and without IBD) are presented in Table 2. Thirty-six IBD patients (42.86%) and eight household members (16.33) needed hospitalization (P = 0.002), with 7 (8.33%) and 2 (4.08%) patients admitted to the ICU, respectively (P = 0.484). While four IBD patients (4.76%) and two household members (4.08%) were intubated, only one patient died in each group.
Univariable analysis showed a significant association between the primary outcome of hospitalization and older age, number of comorbidities more than two, presence of IBD, and new GI symptoms (Table 2). Multivariable logistic regression of the overall study population showed a significant association between age and presence of IBD with hospitalization (Table 3). While each additional year was associated with 6% higher chance of hospitalization (95% CI: 1.03-1.10, P < 0.001), the presence of IBD increased the odds of hospital admission by 5.70 times (95% CI: 2.02-16.07, P = 0.001). Table 4 illustrates the distribution of different outcomes by IBD characteristics and medications. In the univariable analysis, 5-ASA usage was associated with more frequent hospitalization (P < 0.001). There was no significant association between other IBD medications (i.e., anti-TNF monotherapy, immune modulator monotherapy, combination therapy, and corticosteroids) and hospitalization among patients with IBD.
The result of the subgroup analysis among patients with IBD is summarized in Table 5. When we analyzed the association of hospitalization and patients' characteristics, the multivariable logistic regression revealed a significant association between age, new GI symptoms, and 5-ASA usage with the primary outcome of hospitalization among these patients (n = 84). While age was associated with 1.14 times higher odds of hospitalization (95% CI:   and multiple adverse outcomes such as ICU admission, intubation, and death. There was no significant difference between the frequency of severe cases in patients with and without IBD. Furthermore, older age, more than two comorbidities, presence of new GI symptoms and 5-ASA usage were associated with more frequent severe outcome (any case of being admitted to the ICU, intubated, or deceased) (Tables 2 and 4).

Discussion
In this study, we described the clinical characteristics and outcomes in COVID-19 patients with IBD and non-IBD household members. Since IBD patients and their household members had similar exposures, we investigated the risk factors of unfavorable outcomes by comparing these two groups. While hospitalization due to COVID-19 was significantly more frequent in IBD patients compared with non-IBD household members, there was no significant difference between the frequency of severe cases in patients with and without IBD. We observed that age, presence of new GI symptoms, and 5-ASA usage were associated with greater odds of hospitalization among patients with IBD. Anti-TNF use was not significantly related to hospitalization or severe outcomes among patients with IBD. Hospitalization due to COVID-19 was significantly higher among patients with IBD in our study. As far as we know, this is the first time that the hospitalization frequency of patients with IBD is compared with non-IBD household members with similar exposures. This increased risk of hospitalization among IBD patients could be due to worsening of the IBD course or could be related to the COVID-19 course in IBD patients. Also, the baseline activity of IBD could have influenced these results; it is necessary to investigate this factor in future studies and further studies are needed to investigate causality. The 36% hospitalization and 11.2% severe cases with a 1.2% case fatality rate among patients with IBD is relatively similar to previous studies. Brenner et al reported 31% hospitalization, 7% severe COVID-19 and 3% case fatality rate among 525 cases of patients with IBD from 33 countries. 17 A study on outcomes of COVID-19 in 79 patients with IBD in Italy by Bezzio et al showed 28% hospitalization, 3% endotracheal intubation, and 8% death. 16 Also, Higgins et al mentioned that among 239 reported cases, 3% experienced ICU admission. 14 In our study, there was no significant difference between the frequency of severe cases in patients with and without IBD. This finding is similar to the study by Lukin et al, who compared outcomes in COVID-19 patients with or without IBD using a matched cohort design. They reported a severity rate of 24% in IBD patients which was not significantly different from 35% among patients without IBD. 27 Differences in the percentage of severe cases can be attributed to the discrepancy of definitions for disease severity across studies. Furthermore, due to the small event number of severe cases in our study, larger studies are necessary to compare the severity of COVID-19 between patients with and without IBD. Although dyspnea and cough were significantly more common among IBD patients, this could be attributed to recall bias as IBD patients provided information about their household members in our study.
Although Bezzio et al found a significant association between UC diagnosis and IBD activity and COVID-19 pneumonia 16 and Lukin et al reported an association between diagnosis of UC (in comparison to CD) and emergency visit or admission (adjusted odds ratio = 12.7), 27 we found no significant relationship between these variables and COVID-19 outcomes. This finding could be due to the limited number of patients with CD and severe outcomes in our study.
Age was associated with an increased risk of hospitalization and severe outcomes, and this is consistent with previous works. Based on the study by Brenner and Ungaro et al, increasing age is related to 1.04 higher odds of severe COVID-19. 17 Also, Bezzio et al reported a significant association between age over 65 years and COVID-19 pneumonia. 16 We found that the presence of more than two comorbidities was not significantly associated with hospitalization in multivariable analysis, whereas Brenner et al reported 2.9 higher adjusted odds of severe outcomes in patients with more than two comorbidities. This discrepancy may be due to our smaller sample size.
Sixty percent of patients with IBD experienced new GI symptoms, which was double the figure for patients without IBD. This fits well with the results of Lukin et al who found GI symptoms such as diarrhea and abdominal pain to be more frequent among patients with IBD (45 versus 19%, P < 0.001 and 20% versus 5%, P = 0.001, respectively). 27 Also, new GI symptoms were significantly associated with hospitalization and severe outcomes. This is inconsistent with the findings of Lukin et al, who reported no association between emergency visit or admission and GI symptoms. 27 This discrepancy could be attributed to reporting bias in our study or the GI symptoms could be interpreted as markers of IBD activity. Further studies are needed to explore the association between GI symptoms and IBD activity. Also, future studies should explore other possible explanations such as the effect of COVID-19 on IBD or the intestinal susceptibility of SARS-CoV-2 infection in each group. 5-ASA usage was significantly associated with hospitalization among IBD patients, and there was a positive association between 5-ASA and severe outcomes. This is consistent with previous studies that reported sulfasalazine or 5-ASA use related to 3.1 higher odds of severe COVID-19 (95% CI: 1.3-7.7). 17 Ungaro et al reported a 3.52-time higher risk of severe COVID-19 for IBD patients with mesalamine/sulfasalazine usage, in comparison with anti-TNF monotherapy. 28 This finding could be due to under-treatment of IBD and patients having active disease or an effect of the treatment itself. Future studies are needed to clarify this association.
According to our results, there was no significant association between anti-TNF use and hospitalization among patients with IBD. This result is in line with previous studies. 17,27,29 Furthermore, Ungaro et al have reported the possibility of a relative protective effect for anti-TNF usage among patients with IBD. 28 Lukin et al reported numerically higher emergency visits/ hospitalizations among patients with IBD who did not receive biologic medications (P = 0.197). 27 Although corticosteroid use was not associated with hospitalization or severe outcome in our study, this result should be interpreted with caution due to the small number of patients with prednisolone usage in our study population. This is consistent with the study by Bezzio et al, 16 but is in contrast with the findings of Brenner et al 17 Akiyama et al. reported increased risk of severe COVID-19 among patients with autoimmune disease who used glucocorticoids in comparison with anti-TNF monotherapy. 30 The major strengths of our study include the use of a nationwide IRCC cohort for approaching patients with IBD and comparing them with their non-IBD household members with COVID-19. Also, we followed patients for four weeks, and the expert committee confirmed COVID-19 by individual chart review.
It is plausible that some limitations may have influenced the results obtained. To begin with, we enrolled patients who were able to pick up the phone and due to the limited sample size, we enrolled patients with IBD who have called the hotline; these could cause a selection bias. Furthermore, since the IBD patients answered the questions about themselves and their household members, this could lead to a reporting bias. Also, the investigation of rare risk factors, such as Janus kinase inhibitor usage, was not possible. An additional possible source of error is the underestimation of the disease as a result of silent carriers and our diagnostic limitations. Moreover, the household spread of COVID-19 which was low in this cohort could be a result of not testing for COVID-19 and silent carriers.
In conclusion, household spread of COVID-19 was not common in this cohort. New GI symptoms were more common in IBD patients than non-IBD household members. Hospitalization was significantly higher among patients with IBD in our study, and the evidence from this study suggests age, presence of new GI symptoms, and 5-ASA usage as risk factors of hospitalization among patients with IBD. There was no significant difference between the frequency of severe cases in patients with and without IBD. Finally, anti-TNF use was not significantly related to hospitalization or severe outcomes among patients with IBD. Further work needs to be done to establish whether IBD medications are associated with hospitalization or severe outcomes, especially in comparison to patients without IBD.