SARS-CoV-2 Testing, Prevalence, and Predictors of COVID-19 in Patients with Inflammatory Bowel Disease in Northern California

Abstract

Coronavirus disease 2019 (COVID-19), caused by the novel betacoronavirus severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2), is an unprecedented global pandemic.1Rothan H.A. et al.J Autoimmun. 2020; 109: 102433Crossref PubMed Scopus (2891) Google Scholar Susceptibility to COVID-19 is a concern among patients with inflammatory bowel disease (IBD) who are at increased risk of infection due to immunosuppressive therapy. The receptor angiotensin-converting enzyme (ACE) 2, which mediates SARS-CoV-2 entry into cells, is upregulated in IBD2Garg M. et al.J Renin Angiotensin Aldosterone Syst. 2015; 16: 559-569Crossref PubMed Scopus (53) Google Scholar and may therefore increase host susceptibility. International cohorts have reported no increased risk of COVID-19 in patients with IBD3Mao R. et al.Lancet Gastroenterol Hepatol. 2020; 5: 426-428Abstract Full Text Full Text PDF PubMed Scopus (230) Google Scholar,4Norsa L. et al.Gastroenterology. 2020; (Apr 2)PubMed Google Scholar; however, these studies do not report the prevalence of SARS-CoV-2 testing and COVID-19 in patients with IBD. Our institution was among the first to initiate large-scale SARS-CoV-2 RNA testing in northern California. We characterized the prevalence and clinical predictors of COVID-19 in patients with IBD. We performed a retrospective analysis of consecutive patients whose SARS-CoV-2 testing was performed at Stanford between March 04, 2020, and April 14, 2020. California counties tested, institutional testing eligibility, and performance are described in our Supplementary Methods. Our study was approved by the Stanford Institutional Review Board (Protocol 55975). We included all patients with a diagnosis of Crohn’s disease (K50.xx), ulcerative colitis (K51.xx), and indeterminate colitis (K52.3) who underwent testing. We collected data including demographics, IBD characteristics (subtype, location, phenotype, disease activity), comorbid conditions, reasons for testing, symptoms, medications, and outcomes. We calculated prevalence of IBD among all patients tested and the prevalence of COVID-19 among patients with IBD. We performed univariate and multivariate logistic regression using the firthlogit method to determine predictors of COVID-19 in patients with IBD.5Firth D. et al.Biometrika. 1993; : 27-38Crossref Scopus (2535) Google Scholar Our statistical analysis was performed with Statistics/Data Analysis (Stata/IC 15.1 for Windows; StataCorp, College Station, TX) and described in detail in our Supplementary Methods. From March 4, 2020, to April 14, 2020, 14,235 individuals were tested for SARS-CoV-2 at our institution with 8.2% (1160 of 14,235) testing positive. Among the tested patients, the prevalence of IBD was 1.2% (168 of 14,235). Table 1 summarizes the baseline characteristics of patients with IBD who underwent testing; 51.2% had ulcerative colitis, 39.3% had Crohn’s disease, and 9.8% had indeterminate colitis. Of patients with IBD, 16.7% had active disease; 91.7% were symptomatic suggestive of COVID-19, 3.6% were asymptomatic but had a positive travel history, and 4.8% were asymptomatic but had direct exposure to a patient with COVID-19. Common presenting symptoms included cough (63.1%), sore throat (41.1%), dyspnea (37.5%), fever (35.7%), and body pain (32.1%). Gastrointestinal symptoms were present in 19.1% of patients with IBD; diarrhea (15.5%), abdominal pain (13.1%), and nausea and vomiting (8.9%) were most common.Table 1Baseline Clinical Characteristics of Patients With IBD Undergoing SARS-CoV-2 TestingClinical variablesAll patients with IBDSARS-CoV-2 RNASARS-CoV-RNAP(N = 168)Negative (n = 163)positive (n = 5)Age, y (SD)47.7 (±16.3)47.0 (±16.0)70.6 (± 4.2)<.001Age >66, n (%)23 (13.7)19 (11.7)4 (80.0)<.001Gender, n (%)Male80 (47.6)78 (47.9)2 (40.0).810Female88 (52.4)85 (52.1)3 (60.0)Ethnicity, n (%)White103 (61.3)99 (60.7)4 (80.0).344Hispanic14 (8.3)14 (8.3)0 (0.0).501Black13 (7.7)12 (7.4)1 (20.0).283Asian29 (17.3)29 (17.8)0 (0.0).309Pacific Islander1 (0.6)1 (0.6)0 (0.0).863Unknown13 (7.7)13 (8.0)0 (0.0).518Reason for SARS-CoV-2 testing, n (%)Symptomatic154 (91.7)149 (91.4)5 (100).501Asymptomatic, travel history6 (3.6)6 (3.7)0 (0.0).667Asymptomatic, exposure8 (4.8)8 (4.9)0 (0.0).405Clinical features, n (%)Fever60 (35.7)57 (35.0)3 (60.0).230Cough106 (63.1)102 (62.6)4 (80.0).380Nasal congestion58 (34.5)55 (33.7)3 (60.0).200Sore throat69 (41.1)67 (41.1)2 (40.0).996Dyspnea63 (37.5)61 (37.4)2 (40.0).874Fatigue43 (25.6)40 (24.5)3 (60.0).067Body pain54 (32.1)51 (31.3)3 (60.0).159Pneumonia10 (6.0)8 (4.9)2 (40.0).131Gastrointestinal symptoms, n (%)32 (19.0)31 (19.0)1 (20.0).935Abdominal pain22 (13.1)21 (12.9)1 (20.0).589Nausea/Vomiting15 (8.9)14 (8.6)1 (20.0).364Diarrhea26 (15.5)26 (16.0)0 (0.0).338Melena1 (0.6)1 (0.6)0 (0.0).862Hematochezia2 (1.2)2 (1.2)0 (0.0).762Hematemesis1 (0.6)1 (0.6)0 (0.0).862Weight loss5 (3.0)5 (3.0)0 (0.0).695Dysphagia3 (1.8)3 (0.9)0 (0.0).762COVID-19 testing setting, n (%)Outpatient105 (62.5)101 (62.0)4 (80.0).352Emergency department43 (25.6)40 (24.5)3 (60.0).068Inpatient23 (13.7)22 (13.5)1 (20.0).893Ulcerative colitis, n (%)Total86 (51.2)83 (50.1)3 (60.0).641E124 (27.9)23 (27.7)1 (33.3)E219 (22.1)18 (21.7)1 (33.3)E332 (47.1)31 (37.3)1 (33.3)Unknown3 (1.8)3 (3.6)0 (0.0)Crohn's disease, n (%)Total66 (39.3)64 (39.3)2 (40.0).931L113 (19.7)13 (20.3)0 (0.0)L214 (21.2)14 (21.9)0 (0.0)L332 (48.5)30 (46.9)2 (100.0)L40 (0.0)0 (0.0)0 (0.0)Unknown3 (4.7)3 (4.7)0 (0.0)Perianal disease, n (%)12 (18.8)12 (18.8)0 (0.0)B143 (25.6)42 (65.6)1 (50.0)B210 (6.0)9 (14.1)1 (50.0)B38 (12.5)8 (12.5)0 (0.0)Unknown1 (1.6)1 (1.6)0 (0.0)Indeterminate IBD, n (%)Total16 (9.8)16 (9.8)0 (0.0).634BMI, kg/m2, n (%)<25.0 (normal or underweight)90 (53.6)89 (54.6)1 (20.0).14625.0–29.9 (overweight)50 (29.8)49 (30.0)1 (20.0).656≥30.0 (obese)30 (17.9)27 (16.6)3 (60.0).011Smoking, n (%)Current10 (6.0)10 (6.1)0 (0.0).988Former26 (15.5)26 (16.0)1 (20.0)Never131 (80.0)127 (77.9)4 (80.0)Alcohol use, n (%)Yes72 (42.9)69 (42.3)3 (60.0).701No100 (57.1)98 (60.1)2 (40.0)Hypertension, n (%)Yes42 (25.0)38 (23.3)4 (80.0)<.001No126 (75.0)125 (76.7)1 (20.0)Diabetes mellitus, n (%)Yes18 (10.7)16 (9.8)2 (40.0).029No150 (89.3)147 (90.2)3 (60.0)Medications, n (%)ACE inhibitor13 (7.7)10 (6.1)3 (60.0)<.001ARB10 (6.0)10 (6.1)0 (0.0).574PPI33 (19.6)33 (20.2)0 (0.0).271H2 Blocker19 (11.3)18 (11.0)1 (20.0).513Steroids34 (20.2)33 (20.2)1 (20.0).9845-ASA58 (34.5)54 (33.1)4 (80.0).0256MP/Azathioprine9 (5.4)8 (4.9)1 (20.0).131Methotrexate6 (3.6)6 (3.7)0 (0.0).667Anti-TNF agent, no. (%)34 (20.2)33 (20.2)1 (20.0).984Vedolizumab10 (6.0)10 (6.1)0 (0.0).574Ustekinumab4 (2.4)4 (2.5)0 (0.0).727Tofacitinib, no (%)0 (0.0)0 (0.0)0 (0.0)N/AAntiplatelets11 (6.5)10 (6.1)1 (20.0).205Anticoagulant11 (6.5)10 (6.1)1 (20.0).205NSAIDs20 (11.9)20 (12.2)0 (0.0).4125-ASA, mesalamine; ARB, angiotensin receptor blocker; B1, nonstricturing, nonpenetrating CD; B2, stricturing CD; B3, penetrating CD; BMI, body mass index; CD, Crohn's disease; E1, distal UC; E2, left-sided UC; E3, extensive UC; L1, ileal CD, L2, colonic CD, L3, ileocolonic CD; NSAID, nonsteroidal anti-inflammatory drug; PPI, proton pump inhibitor; TNF, tumor necrosis factor; UC, ulcerative colitis. Open table in a new tab 5-ASA, mesalamine; ARB, angiotensin receptor blocker; B1, nonstricturing, nonpenetrating CD; B2, stricturing CD; B3, penetrating CD; BMI, body mass index; CD, Crohn's disease; E1, distal UC; E2, left-sided UC; E3, extensive UC; L1, ileal CD, L2, colonic CD, L3, ileocolonic CD; NSAID, nonsteroidal anti-inflammatory drug; PPI, proton pump inhibitor; TNF, tumor necrosis factor; UC, ulcerative colitis. Among 168 patients with IBD tested, the prevalence of COVID-19 was 3.0% (5 of 168). Patients with IBD with COVID-19 were older (70.6 years vs 47 years, P < .001), more obese (60.0% vs 16.6%, P = .011), and more likely to have hypertension (80.0% vs 23.3%, P < .001) and diabetes mellitus (40.0% vs 9.8%, P = .029). Patients with IBD with COVID-19 were more likely to use ACE inhibitors (60.0% vs 6.1%, P < .001) and mesalamine (80.0% vs 33.1%, P = .025). In univariate analysis (Supplementary Table 1), age >66 years (odds ratio [OR] 31.37, P = .003), obesity (BMI ≥30) (OR 7.83, P = .011), hypertension (OR 13.58, P = .021), and ACE inhibitor use (OR 23.70, P = .001) were associated with increased risk of COVID-19 among patients with IBD. Our multivariate logistic regression model, which included age >66 years, obesity, hypertension, and ACE inhibitor use as covariates, showed that age >66 years was independently associated with increased risk (OR 21.30, P = .022) of COVID-19. Clinical outcomes of patients with IBD with COVID-19 are summarized in Supplementary Table 2. Four patients with IBD had a mild course, whereas 1 patient (Patient 3) developed pneumonia and acute respiratory distress syndrome and died despite aggressive interventions. To our knowledge, this is the first study to evaluate the prevalence of SARS-CoV-2 testing and COVID-19 in patients with IBD in a US cohort. The prevalence of IBD among patients undergoing SARS-CoV-2 testing is 1.2%, which is comparable to the prevalence of IBD (1.3%) in the US adult population.6Cdc.gov. 2020.https://www.cdc.gov/ibd/data-statistics.htmDate accessed: April 24, 2020Google Scholar Our COVID-19 positivity rate of 3% in patients with IBD is comparable to the population-weighted prevalence of SARS-CoV-2–positive serology in Santa Clara county at 2.8%.7Bendavid E. et al.medRxiv. 2020; PubMed Google Scholar Our data suggest that patients with IBD are not disproportionately being tested more, nor do they have a higher rate of SARS-CoV-2 positivity compared with the background population in northern California. One explanation is that increased ACE 2 expression may not mediate SARS-CoV-2 susceptibility in patients with IBD. Another possibility is that immunosuppressive medications in patients with IBD may attenuate viral-induced respiratory inflammation leading to an asymptomatic or mild COVID-19 course in patients with IBD who subsequently do not seek testing. Our study also demonstrates that patients older than 66 years are at increased risk of COVID-19. Our results are consistent with a prior retrospective study from China that demonstrated that older age is an independent predictor of COVID-19.8Shi Y. et al.Crit Care. 2020; 24: 1-4Crossref PubMed Scopus (27) Google Scholar The exact mechanisms underlying susceptibility to COVID-19 in elderly patients are unclear and warrant further investigation. Our study has several strengths. First, our study provides novel epidemiological data that can inform patients with IBD and clinicians. Currently, there are no published reports estimating the prevalence of COVID-19 among patients with IBD in the United States. Second, we identified predictors of COVID-19 among patients with IBD, highlighting the increased susceptibility of COVID-19 with older age. Third, our study included patients from a large geographic area encompassing a diverse patient population. Our study has several limitations. First, our study was observational and cannot establish causation or account for unmeasured confounders. Second, we were unable to assess the predictors of COVID-19 morbidity and mortality with our small sample size and low event rate. A significantly larger sample size is needed to further clarify predictors of COVID-19 outcomes. Third, our study reflects testing performed by a single center and may not be generalizable to other institutions. In summary, our results provide much needed epidemiological data and reassurance that COVID-19 rates in patients with IBD may be comparable to the general population. Age older than 66 years was a strong independent predictor of COVID-19 among patients with IBD. We are extremely grateful to our Stanford IBD clinical and research team members Gayathri Swaminathan, PhD, David Limsui, MD, Sidhartha Sinha, MD, Kian Keyashian, MD, Sarah E. Streett, MD, and Chris Cartwright, MD for their critiques and helpful feedback in drafting the manuscript. John Gubatan, MD (Conceptualization: Lead; Data curation: Lead; Formal analysis: Lead; Investigation: Lead; Methodology: Lead; Software: Lead; Writing – original draft: Lead; Writing – review & editing: Lead). Steven Levitte, MD, PhD (Data curation: Supporting; Writing – original draft: Supporting; Writing – review & editing: Supporting). Tatiana Balabanis, BA (Data curation: Supporting). Akshar Patel, BA (Data curation: Supporting). Arpita Sharma, PhD (Data curation: Supporting). Aida Habtezion, MD, MSc (Conceptualization: Equal; Formal analysis: Equal; Supervision: Equal; Writing – review & editing: Supporting). Patients resided in several northern California counties, including Santa Clara, San Mateo, Santa Cruz, San Francisco, and Alameda. Patients were offered SARS-CoV-2 RNA testing at our institution if they had symptoms or findings suggestive of COVID-19 (fevers, cough, dyspnea, pneumonia), had a recent travel history with high COVID-19 cases, or had direct exposure to a patient with COVID-19. All SARS-CoV-2 RNA testing was performed using samples from a nasopharyngeal swab. The clinical sensitivity of the COVID-19 test at our institution is 96% (using repeat testing within 48 hours as a surrogate gold standard and assuming all negatives are false negatives) and clinical specificity approaches 100%. The rate of SARS-CoV-2–positive tests, predictive value of clinical variables on the primary outcome, OR with its 95% confidence interval, and P values were assessed using Statistics/Data Analysis (Stata/IC 15.1 for Windows, StataCorp, College Station, TX). We calculated the prevalence of IBD among patients undergoing SARS-CoV-2 testing by dividing the number of patients with IBD tested by total number of patients tested in our population. The prevalence of SARS-CoV-2 positivity was calculated by dividing the number of patients with IBD with positive SARS-CoV-2 tests over the number of total patients with IBD tested. Dichotomous variables were analyzed for outcomes using the χ2 test or the Fisher exact test where appropriate, and continuous variables were analyzed using t tests if normally distributed, or the Wilcoxon test for non-normal data. Correction for multiple testing was included. All variables were analyzed initially in a univariate fashion to determine their association with COVID-19. P values of factors that showed evidence of an association on COVID-19 (P < .05) then were analyzed on multivariate regression analysis. Rare events may lead to complete separation and problems with convergence in conventional logistic regression models. The Firth method is a general approach to reducing rare event and small-sample bias in maximum likelihood estimation. Because of the small sample size and low event rate for outcome of patients with COVID-19 with IBD, we used the firthlogit penalized maximum likelihood logistic regression in our analysis.Supplementary Table 1Univariate and Multivariate Predictors of COVID-19 Among Patients With IBDClinical variablesUnivariate predictorsMultivariate predictorsOR95% CIPOR95% CIPAge >66 y31.373.33–295.46.00321.301.56–291.00.022Obesity (BMI ≥30)7.831.25–49.12.0111.350.09–21.54.830Hypertension13.581.47–125.15.0213.650.30–45.11.313Diabetes mellitus6.290.98–40.49.053ACE inhibitor use23.703.55–158.44.00110.610.67–168.09.094Mesalamine (5-ASA)8.440.92–77.37.059BMI, body mass index; CI, confidence interval. Open table in a new tab Supplementary Table 2Clinical Characteristics and Outcomes in Patients With IBD With COVID-19PatientDemographicsMontrealDiseaseIBDCOVID-19Mild COVID-19Severe COVID-19numberethnicityclassificationactivitymedicationssymptomsOutptaPatient died of acute respiratory distress syndrome.EDbED = Emergency Department.HospcHosp = Inpatient Hospitalization.ICUdICU = Intensive Care Unit Admission.MVeMV = Mechanical Ventilation.Death168 FCDAPrednisoneFever, CoughYesYesNoNoNoNoWhiteL3, B2IFXFatigue274 MCDR5-ASACoughYesYesNoNoNoNoWhiteL3, B1376 MUCR5-ASAFeverNoNoYesYesYesYesaPatient died of acute respiratory distress syndrome.BlackDyspnea469 FUCR5-ASAFever, CoughYesNoNoNoNoNoWhiteE3Fatigue566 FUCR5-ASACoughYesNoNoNoNoNoWhiteE2AZADyspnea5-ASA, mesalamine; A, active; AZA, Azathioprine; B1, nonstricturing, nonpenetrating CD; B2, stricturing CD; B3, penetrating CD; CD, Crohn's disease; E1, distal UC; E2, left-sided UC; E3, extensive UC; ED, emergency department; F, female; Hosp, hospitalization; IFX, Infliximab; ICU, intensive care unit; L1, ileal CD, L2, colonic CD, L3, ileocolonic CD; M, male; MV, mechanical ventilation; Outpt, outpatient; R, remission; UC, ulcerative colitis.a Patient died of acute respiratory distress syndrome.b ED = Emergency Department.c Hosp = Inpatient Hospitalization.d ICU = Intensive Care Unit Admission.e MV = Mechanical Ventilation. Open table in a new tab BMI, body mass index; CI, confidence interval. 5-ASA, mesalamine; A, active; AZA, Azathioprine; B1, nonstricturing, nonpenetrating CD; B2, stricturing CD; B3, penetrating CD; CD, Crohn's disease; E1, distal UC; E2, left-sided UC; E3, extensive UC; ED, emergency department; F, female; Hosp, hospitalization; IFX, Infliximab; ICU, intensive care unit; L1, ileal CD, L2, colonic CD, L3, ileocolonic CD; M, male; MV, mechanical ventilation; Outpt, outpatient; R, remission; UC, ulcerative colitis. What Is the Incidence of COVID-19 in Patients With IBD in Western Countries?GastroenterologyVol. 160Issue 5PreviewWe read with interest the article by Gubatan et al1 reporting that, among 168 patients with inflammatory bowel disease (IBD) tested in Northern California (Stanford University School of Medicine), the prevalence of coronavirus disease 2019 (COVID-19) was 3.0%, comparable with the population-weighted prevalence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)–positive serology in Santa Clara County at 2.8%. The authors concluded that their results provided much-needed epidemiologic data and reassurance that COVID-19 rates in patients IBD may be comparable with that in the general population. Full-Text PDF