Corticosteroids but not Anti-TNF Are Associated With Increased COVID-19 Complications in Patients With Inflammatory Bowel Disease

Abstract

Background and Aims

Immunosuppressed individuals are at higher risk for COVID-19 complications, yet data in patients with inflammatory bowel disease (IBD) are limited. We evaluated the risk of COVID-19- severe sequelae by medication utilization in a large cohort of patients with IBD.

Methods

We conducted a retrospective cohort study utilizing insurance claims data between August 31, 2019, and August 31, 2021.We included IBD patients identified by diagnosis and treatment codes. Use of IBD medications was defined in the 90 days prior to cohort entry. Study outcomes included COVID-19 hospitalization, mechanical ventilation, and inpatient death. Patients were followed until the outcome of interest, outpatient death, disenrollment, or end of study period. Due to the aggregate nature of available data, we were unable to perform multivariate analyses.

Results

We included 102 986 patients (48 728 CD, 47 592 UC) with a mean age of 53 years; 55% were female. Overall, 412 (0.4%) patients were hospitalized with COVID-19. The incidence of hospitalization was higher in those on corticosteroids (0.6% vs 0.3%; P < .0001; 13.6 per 1000 person-years; 95% confidence interval [CI], 10.8-16.9) and lower in those receiving anti-tumor necrosis factor α therapy (0.2% vs 0.5%; P < .0001; 3.9 per 1000 person-years; 95% CI, 2.7-5.4). Older age was associated with increased hospitalization with COVID-19. Overall, 71 (0.07%) patients required mechanical ventilation and 52 (0.05%) died at the hospital with COVID-19. The proportion requiring mechanical ventilation (1.9% vs 0.05%; P < .0001; 3.9 per 1000 person-years; 95% CI, 2.5-5.9) was higher among users of corticosteroids.

Conclusions

Among patients with IBD, those on corticosteroids had more hospitalizations and mechanical ventilation with COVID-19. Anti-tumor necrosis factor α therapy was associated with a decreased risk of hospitalization. These findings reinforce previous guidance to taper and/or discontinue corticosteroids in IBD.

Introduction

Risk factors for severe COVID-19 complications include immunosuppressant medications (namely corticosteroids),¹ immune-mediated inflammatory diseases,² lack of vaccination,³ older age, and comorbidities.^4,5 Inflammatory bowel diseases (IBDs), including Crohn’s disease (CD) and ulcerative colitis (UC), are chronic immune-mediated inflammatory conditions of the gastrointestinal tract affecting millions of people worldwide.^6-8 In North America, the prevalence of IBD is estimated to be between 400 and 600 cases per 100 000 persons.⁹ Patients with IBD are often treated with immunosuppressive therapies such as corticosteroids, anti-tumor necrosis factor α (anti-TNF) agents, immunomodulators (thiopurines or methotrexate), anti-integrins, JAK inhibitors, and anti-interleukin-12/23 agents. As such, affected individuals may be at increased risk for complications of COVID-19.

Prior work, including the international SECURE-IBD (Secure Epidemiology of Coronavirus Under Research Exclusion for Inflammatory Bowel Disease) registry,¹⁰ has described medication-specific risks of complications from COVID-19. However, these data represent a convenience sample based on voluntary reporting of cases by physicians and may be subject to reporting and other biases. Additionally, this registry does not include a population-based denominator from which to calculate medication specific rates of COVID-19 hospitalization. These data are needed to address patient concerns regarding medication use and inform treatment decisions during this pandemic.¹¹

We therefore aimed to determine whether different medication classes for the treatment of IBD (compared with each other and no medications) impact the risk of developing severe COVID-19 (hospitalization, mechanical ventilation, or inpatient death) using a large population-based cohort. We aimed to calculate rates per 1000 person-years and 95% confidence intervals (CIs) for each outcome of severe COVID-19 by medication class.

Methods

Study Design

We conducted a retrospective cohort including data from 2 large U.S. health plans that participate in the Food and Drug Administration Sentinel System—Anthem and Humana—between August 31, 2019, and August 31, 2021. Cohort entry and measurement of outcomes began on March 1, 2020, the beginning of the COVID-19 pandemic in the United States, and ended with occurrence of a study outcome, outpatient death, loss of medical coverage, or end of available data.

Data Source

We analyzed the outpatient and inpatient diagnosis and procedure data and outpatient pharmacy dispensing data contained in Anthem and Humana. These longitudinal, patient-level databases have been used in previous epidemiological studies of immune-mediated conditions such as IBD.^12,13 The included plans capture a geographically diverse sample. Prior studies have reported these administrative databases to be representative of the national commercially insured population on a variety of demographic measures, including age and sex.¹³

Patient Selection

To be classified as having IBD, the member had to have met 1 of 3 definitions: (1) 2 International Classification of Diseases–Tenth Revision–Clinical Modification (ICD-10-CM) diagnoses of IBD in the prior 183 days; (2) 1 ICD-10-CM diagnosis of IBD and treatment with mesalamine, sulfasalazine, bulsalazide, olsalazine, or vedolizumab as assessed using National Drug Code (NDC) and Healthcare Common Procedure Coding System (HCPCS) codes; or (3) 1 ICD-10-CM diagnosis of IBD and treatment with anti-TNF, ustekinumab, azathioprine, 6-mercaptopurine, or methotrexate as assessed using NDC and HCPCS codes in the absence of an ICD-10-CM diagnosis of rheumatoid arthritis, psoriasis, psoriatic arthritis, pyoderma gangrenosum, hidradenitis suppurativa, or ankylosing spondylitis. The IBD subtype of either UC or CD was determine based on ICD-10-CM diagnoses within the 183-day period. A subtype was assigned if all diagnoses matched a single subtype. If diagnoses were mixed, no subtype was assigned, and the member was considered IBD not otherwise specified for the purposes of this analysis.

Assessment of Exposures (Medication Use)

We categorized exposure to IBD treatment based on evidence of days’ supply in the 90 days prior to cohort entry using NDC and HCPCS. No evidence of treatment for any of the drugs of interest resulted in categorization as no treatment. The primary medications we evaluated included prednisone, methylprednisolone, prednisolone, hydrocortisone, and dexamethasone (systemic corticosteroids); oral budesonide, sulfasalazine, and mesalamine derivatives (aminosalicylate class [5-aminosalicylate]); azathioprine and 6-mercaptopurine (thiopurine class); methotrexate, infliximab, adalimumab, golimumab, and certolizumab pegol (anti-TNF biologic class); vedolizumab and natalizumab (anti-integrin class); tofacitinib (JAK inhibitor class); and ustekinumab (anti-interleukin-12/23 class).

Baseline Characteristics

In addition to age and sex, we assessed recent medical diagnoses associated with IBD that could have affected COVID-19 risk or severity including type 2 diabetes, chronic kidney disease, obesity diagnosis, pregnancy, and chronic obstructive pulmonary disease in the 183 days prior to cohort entry. We also assessed health services utilization and drug utilization in the 183 days prior to cohort entry.

Study Outcomes

We identified the first instance of COVID-19 inpatient visit using the ICD-10-CM codes most frequently associated with COVID-19 in the inpatient setting.¹⁴ Mechanical ventilation was assessed using ICD-10 Procedure Coding System and Current Procedural Terminology (fourth edition) codes. Inpatient death was identified using encounter discharge status or if the member’s death date fell between admission and discharge dates for the inpatient stay.

Analysis Plan

We utilized descriptive statistics to summarize characteristics of IBD patients overall and those hospitalized with COVID-19 with a particular focus on comorbidities and medication. We utilized bivariate analyses to determine characteristics associated with the risk of inpatient hospitalization, ventilator use, or inpatient death. We then determined rates and 95% CIs for COVID-19 inpatient admission and ventilator use per 1000 person-years by IBD medication classes, sex, and age. Rates and 95% CI for inpatient death could not be calculated due to low numbers of events.

The study protocol was approved by the institutional review board of the University of North Carolina. This analysis was completed using version 10.2.1 of the Sentinel Query Request Package with ad hoc programming.¹⁵

Results

The study population included 102 986 patients (48 728 CD and 47 592 UC) with a mean age of 53 years; 55% were female. Approximately one-quarter of the population were on no treatment. Anti-TNF therapy with or without concurrent immunomodulator was the most common immunosuppressive treatment (19%). Characteristics of the cohort overall and stratified by CD or UC are shown in Table 1.

Overall, 419 (0.4%) patients were hospitalized for COVID-19 infection. Risk factors for hospitalization included older age (P < .001), multiple comorbidities such as obesity (P < .001), and systemic corticosteroids (P < .001). Anti-TNF biologics and 5-aminosalicylate/sulfasalazine were associated with a reduced risk of hospitalization (P < .001) (Table 2). When stratified by CD and UC, we found a continued risk for hospitalization associated systemic corticosteroids and an inverse association with anti-TNF use (Supplemental Tables 1 and 2). The inverse association with anti-TNF use persisted in a subanalysis including only those with no systemic corticosteroid use (16% anti-TNF use among those not hospitalized, 7% among those hospitalized; P < .001). Sex, age, and comorbidities were comparable among corticosteroid users as compared with those using other classes of immunosuppressive medications. However, rates of smoking were higher (23%) for corticosteroid users (Supplemental Table 3).

The COVID-19 inpatient admission rate per 1000 person-years is shown in Figure 1. Rates of inpatient admission were highest for systemic corticosteroids (13.6 per 1000 person-years; 95% CI, 10.8-17.0), with a greater rate of hospitalization when compared with aminosalicylates (6.6 per 1000 person-years; 95% CI, 5.2-8.3) or anti-TNF agents (3.9 per 1000 person-years; 95% CI, 2.7-5.4). Similar effects were seen after stratifying by CD or UC diagnosis and by sex (data not shown). When stratified by age group, rates of inpatient admissions were substantially higher for the group ≥65 years of age for each medication class (Figure 1). For corticosteroid use, the risk was of inpatient admission was particularly notable in the oldest age strata (Supplemental Figure 1).

A total of 72 patients required inpatient ventilator use and 52 patients died from COVID-19 during hospitalization. The majority of those requiring ventilator use were ≥65 years of age (72%) and male (58%). Among those with COVID-19–related death (n = 52), 85% were ≥65 years of age and 44% were male.

The COVID-19–associated inpatient ventilation rate per 1000 person-years is shown in Figure 2. Similarly, systemic corticosteroid use was associated with a higher rate of ventilation (3.9 per 1000 person-years; 95% CI, 2.5-5.9) as compared with anti-TNF use (0.6 per 1000 person-years; 95% CI, 0.2-1.3). Rates per 1000 person-years of inpatient death with COVID-19 were not calculated due to sample size.

Discussion

In this retrospective cohort study of over 100 000 patients with IBD, we found that systemic corticosteroids were associated with higher rates of inpatient admission and ventilator use associated with COVID-19. This may be related to the effects of corticosteroids themselves or to the active disease that was the trigger for initiation of corticosteroids. Importantly, the most common biologics used in this national sample of IBD patients, including anti-TNF agents, vedolizumab, and ustekinumab, were not associated with an increased risk of severe COVID-19. This may be related to a number of factors, including differences in IBD severity or control, or possibly related to a protective effect in the case of anti-TNF. The mechanism of SARS-CoV-2 hyperinflammation is not certain. High serum anti-TNF concentrations in COVID-19 patients have been associated with worse outcomes.¹⁶ Therefore, anti-TNF therapies may block this hyperimmune response seen with COVID-19.¹⁷ A separate cohort of >6000 patients with immune-mediated inflammatory disorders demonstrated the lowest risk of hospitalization or death from COVID-19 with anti-TNF monotherapy as compared with other therapies.¹⁷

These results are both timely and relevant, given the concerns in the IBD patient community surrounding medication utilization and risks of complications from COVID-19.¹¹ A recent survey found that active disease, biologic treatment, and use of corticosteroids in the last 3 months were perceived by the patients as high-risk features for increased risk of COVID-19 infection and more severe disease.¹⁸ As fears of infectious complications surrounding immunosuppressive medication use can lead to cessation or interruption in therapy, these data inform a critical gap for patients.

A number of observational cohort studies have been conducted to investigate adverse events associated with COVID-19 in patients with IBD including development of pneumonia, hospitalization, need for ventilator support, and death. In an Italian cohort of 937 individuals testing positive for COVID-19, 13.7% were asymptomatic, 70.8% had a favorable course, and 15.5% had moderate or severe COVID-19. Factors associated with adverse outcomes included obesity, active disease, and comorbidities. Interestingly, no single IBD medication was associated with worse COVID-19 outcomes, although this may have been limited by sample size.¹⁹ A recent meta-analysis evaluated the safety of therapies for IBD in relation to severe COVID-19 outcomes. Corticosteroids were associated with increased risk of severe COVID-19 (relative risk [RR], 1.91; 95% CI, 1.25-2.91) as were mesalamine therapies (RR, 1.50; 95% CI, 1.17-1.93).¹¹ Corticosteroids increased the risk of intensive care unit admission but not mortality in this study. Both ustekinumab (RR, 0.55; 95% CI 0.43-0.72) and anti-TNF (RR, 0.47; 95% CI, 0.40-0.54) were associated with a decreased risk of severe COVID-19. However, in this study, patient-level data were lacking, and insufficient data existed for meta-regression analyses to adjust for confounding.¹¹ A separate meta-analysis found similar results.¹⁹ For all of these earlier observational studies, the possibility of residual confounding remains.

A strength to our retrospective cohort study is the large sample size, which permits drug-specific comparisons. Additionally, we were able to include complete capture of hospitalizations and other outcomes across health systems. Due to the large sample and ability to define a source population, we were able to calculate rates per 1000 person-years of inpatient COVID-19 hospitalization by specific medications. There were also limitations to this study. We used administrative data, and thus there is the possibility of misclassification of exposure or outcome. However, we did use a validated algorithm for defining our IBD cohort with a sensitivity of over 95% and a positive predictive value ranging from 81.2% to 94.1%.²⁰ While there is a possibility of misclassification bias, this would most likely be nondifferential by drug exposure class. Claims data accurately depict prescriptions of medications, although we do not have specific data on adherence to oral or home injection medications. With infused biologics, these encounters were associated with delivery of the therapy, allowing for more certainty as to adherence. We do not have access to socioeconomic factors such as income, although this cohort does represent a commercially insured population. Insurance is one factor that improves access to care in the United States. In addition, we also do not have access to clinical data of symptoms or endoscopic severity, and thus we were unable to stratify by IBD severity. While we do not have body mass index, we were able to assess for obesity and other comorbidities. These factors have previously been associated with increased complications of COVID-19. Rates of obesity did not differ by drug class (Supplemental Table 3). Due to the aggregate nature of available data, we were unable to perform multivariate analyses, and therefore the results are largely unadjusted. To account for this, we did a number of analyses within various strata including age and sex. We also stratified demographics and comorbidities by medication class and found comparable ages and comorbidity rates regardless of specific IBD therapeutic agent. However, higher rates of smoking were found among corticosteroid users. Smoking may also contribute to the higher risk of severe COVID-19 that we found in this group.²¹ Although were able to capture hospitalized outcomes well within this cohort, many sources of testing for COVID-19 infection are not included in claims data. Therefore, we were not able to investigate risk factors for the development of COVID-19 infection. Arguably, we focused on those outcomes most relevant to patients’ concerns, severe complications of COVID-19 requiring inpatient admission. The time period examined in this study involved an earlier COVID-19 variant, and extrapolating these data to subsequent variants such as omicron may not be appropriate. Finally, we did not have access to vaccination data, as vaccines are provided free of charge in the United States and often not included in health insurance databases. Additionally, the time period of study was predominantly before vaccine availability. Vaccination for COVID-19 has been shown to be safe and effective in patients with IBD.^22-26

Conclusions

There are a number of implications to this study, namely that controlling disease activity and utilizing steroid-sparing therapy in the management of IBD remains an important tenet of patient care during the COVID-19 pandemic. These data show that corticosteroid use, particularly in older individuals, is associated with a much higher risk of severe COVID-19 than other classes of medications, such as biologic therapies. Therefore, optimizing noncorticosteroid therapies to induce and maintain remission should be prioritized in IBD management. In particular, sharing these data with patients and providers may assuage anxieties and fears surrounding continuation of their maintenance medications for IBD.

Acknowledgments

All statements in this report, including its findings and conclusions, are solely those of the authors and do not necessarily represent the views of the Patient-Centered Outcomes Research Institute (PCORI), its Board of Governors or Methodology Committee.

Author Contribution

M.D.L.: study design, interpretation of analysis, drafting and editing of the manuscript, L.P.: analysis and interpretation of data, editing of the manuscript; J.D.L.: study design, interpretation of analysis, drafting and editing of the manuscript; K.H.: analysis and interpretation of data, editing of the manuscript; Q.M.: data extraction, analysis, editing of the manuscript; S.A.: study design, drafting and editing of the manuscript; J.B.: study design, drafting and editing of the manuscript, JD: study design, drafting and editing of the manuscript; S.T., L.H., A.W.: study design, analysis and interpretation of data, drafting and editing of the manuscript; M.D.K.: study design, interpretation of analysis, drafting and editing of the manuscript. Guarantor of the article: M.D.L. All authors approved the final version of the manuscript prior to submission.

Funding

This work was supported through a Patient-Centered Outcomes Research Institute (PCORI) Award (PaCR2017C2-8172-IC).

Conflicts of Interest

M.D.K. has served as a consultant for AbbVie, BMS, Calibr, Genentech, Gilead, Janssen, Pfizer, Eli Lilly, Prometheus, Roche, Salix, Takeda, Target Pharmasolutions, UCB, Valeant, and Theravance; is a shareholder in Johnson & Johnson; and has received research support from Pfizer, Takeda, Janssen, AbbVie, Lilly, Genentech, Boehringer Ingelheim, Bristol Myers Squibb, Celltrion, and Arena Pharmeceuticals. L.P. is an employee of HealthCore. J.D.L. has served as a consultant or served on an advisory board for Eli Lilly, Samsung Bioepis, UCB, Bristol-Myers Squibb, Nestlé Health Science, Merck, Celgene, Janssen Pharmaceuticals, Bridge Biotherapeutics, Entasis Therapeutics, AbbVie, Pfizer, Gilead, Arena Pharmaceuticals, Protagonist Therapeutics, Amgen, and Scipher Medicine; received research funding from Nestlé Health Science, Takeda, Janssen Pharmaceuticals, and AbbVie; performed legal work on behalf of generic manufacturers of ranitidine, including L. Perrigo Company, Glenmark Pharmaceuticals, Amneal Pharmaceuticals LLC, Aurobindo Pharma USA, Dr. Reddy’s Laboratories, Novitium Pharma, Ranbaxy Inc, Sun Pharmaceutical Industries, Strides Pharma, and Wockhardt USA LLC. K.H. worked for HealthCore during the preparation of this manuscript and is now an employee of Janssen Pharmaceuticals. Q.M. is an employee of Humana. S.T. has served as a consultant for Pfizer and Merck for unrelated work. All other authors have no conflicts to disclose.

Data Availability

Access to study materials (output and tables of deidentified data) can be provided upon request to researchers. Raw data will not be made available, as this study was conducted under specific licensing.

References