Introduction

Note: Some of the results of this study have been previously reported in the form of an abstract presented at the American Thoracic Society Conference 2021.

Severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) has caused, to date, 41 million cases and 667,244 deaths in the United States.¹ Coronavirus disease 2019 (COVID-19), the infectious disease caused by SARS-CoV-2, has become the leading cause of death in the United States.²

Chronic obstructive pulmonary disease (COPD) is the fourth leading cause of death in the United States and affects 16 million Americans.³ During the COVID-19 pandemic, the association of COPD with adverse outcomes due to COVID-19 has been controversial. Worldwide, studies of patients with COVID-19 have reported COPD to be a risk factor for greater health care utilization,⁴ with increased risk of hospitalization, intensive care unit (ICU) admission, and death,^5-12 while others have not shown this association.^13-16 The association between COPD and adverse outcomes from COVID-19 remains disputed.

The prevalence of COPD among patients with COVID-19 varies^8,9,14,17,18 between 1.5% and 18%. This is likely an underestimation, as a SARS-CoV-2 infection can range in presentation from asymptomatic to severe disease. Severe COVID-19 may lead to hospitalization, ICU admission, non-invasive and invasive mechanical ventilation use, and death.^19-24 Given the propensity for patients with COPD to suffer acute exacerbations induced by viral infections, such patients may be particularly vulnerable to adverse outcomes associated with COVID-19.²⁵

The prevalence of COPD varies globally,²⁶ therefore, the reported potential association of COPD as a risk factor for adverse outcomes in COVID-19 may vary by country or national health care system. Consequently, it is important to explore the characteristics and outcomes of patients hospitalized in the United States with COVID-19 and COPD. To evaluate COPD as an independent risk factor for mortality in patients with COVID-19, we conducted a large, nationally representative retrospective cohort study of U.S. patients hospitalized with COVID-19. We hypothesized that, among hospitalized patients with COVID-19, those with COPD would have an increased risk of inpatient mortality.

Methods

Data Source

In this retrospective cohort study, we used Optum’s longitudinal COVID-19 electronic health record (EHR) database of more than 90 million patients across multiple hospital networks from all regions in the United States. The database contains de-identified inpatient and ambulatory encounter-level information, as well as procedure, prescription, and medication administration data. The University of Texas Medical Branch Institutional Review Board approved this study (IRB# is 20-0180). Written informed consent was not required due to the de-identified nature of the patient data.

Cohort

The study cohort included patients diagnosed with COVID-19 between February 10, 2020, and November 10, 2020, who were hospitalized within 14 days of diagnosis. COVID-19 was identified by a positive laboratory test for SARS-CoV-2 or by the International Classification for Diseases, 10th revision, Clinical Modification (ICD-10-CM) diagnosis code U07.1 (See Table S1 in the online data supplement). Patients younger than 40 years at the time of diagnosis were excluded due to the low prevalence of COPD in this population²⁷ (Figure S1 in the online data supplement).

Variables

The primary outcome was inpatient mortality among patients with COVID-19-associated hospitalization. The main independent variable of interest was COPD and was defined as having experienced ≥ 1 inpatient or ≥2 outpatient visits for COPD in the 1 year before the COVID-19 diagnosis (See Table S2 in the online data supplement).

We collected information on patient demographics as well as clinical and medication history. Comorbidities present before the COVID-19 diagnosis were identified using ICD-10-CM diagnosis codes. Convalescent plasma and medications administered during hospitalization (remdesivir, systemic corticosteroids) were identified from National Drug Codes or procedure codes. Inhaled corticosteroid use was defined as having at least 1 prescription in the 12 months preceding the COVID-19 diagnosis. For body mass index (BMI) and insurance status, when multiple observations were available, we recorded the value at the date closest to the COVID-19 diagnosis.

Statistical Analysis

Patient and clinical characteristics were summarized as frequencies and percentages or means±standard deviations and compared with chi-square statistics or t-tests as appropriate. To determine the effect of COPD on inpatient mortality, we fit a logistic regression model with COPD as the primary predictor, adjusted for covariates. We tested a priori interactions between COPD and sex, COPD and age, COPD and race, and COPD and inhaled corticosteroids. For significant interaction terms, we stratified our cohort in subgroups and examined the effect of COPD on inpatient mortality within each stratum with separate multivariate models. All analyses were performed with SAS 9.4 (SAS, Inc., Cary, North Carolina).

Results

Demographics and Patient Characteristics

During the study period, 31,526 patients hospitalized with COVID-19 were identified and baseline characteristics of the cohort are presented in Table 1. Patients with COPD comprised 15.09% of the cohort (4758 patients), which consisted of mostly White (68.6%) males (52.1 %) with a BMI < 30kg/m² (46%) and mean age of 72±11.2 years. A greater percentage of patients with COPD were admitted to the ICU (COPD 28.8% versus non-COPD 22.2%, p<0.0001) and received mechanical ventilation (COPD 20.2% versus non-COPD 13.8%, p<0.0001), systemic steroids (COPD 56.3% versus non-COPD 43.7%, p<0.0001), convalescent plasma (COPD 4.5% versus non-COPD 3.7%, p=0.0053), and remdesivir (COPD 12.9% versus non-COPD 11.5%, p=0.0051). Similarly, a greater percentage of patients with COPD received a palliative care consultation (COPD 20.3% versus non-COPD 14.3%, p<0.0001) (Table 2). In addition, more patients with COPD were discharged to a non-home setting compared to patients without COPD. In patients with COPD, the most common non-home setting discharge destination was a skilled nursing facility (18.2%), followed by hospice (4.6%) (Table S3 in the online data supplement).

Effect of COPD on COVID-19 Inpatient Mortality

Overall hospital mortality was 9.61% for hospitalized patients with COVID-19. Mortality was significantly higher in patients with COPD than in patients without COPD, 14.02% and 8.83%, respectively, p<0.0001 (Table 2). In univariate analysis, patients with COPD had 68% greater odds of death due to COVID-19 than patients without COPD (odds ratio [OR] 1.68; 95% confidence interval [CI] 1.54 to 1.84). After adjusting for clinical and demographic factors, we observed a modest reduction in the odds of mortality in patents with COVID-19 and COPD (OR 1.33; 95% CI 1.18 to1.50) (See Table 3 ). We found significant interactions between COPD and sex and COPD and age, but not between COPD and race or COPD and inhaled corticosteroid (ICS) use. Specifically, the increased mortality risk associated with COPD was observed among female (OR 1.62; 95% CI 1.36 to 1.95) but not male patients (OR 1.14; 95% CI 0.97 to 1.34); and in patients aged 40 to 64 (OR 1.42; 95% CI 1.07 to 1.90) and 65 to 79 (OR 1.48; 95% CI 1.23 to 1.78) years, but not in patients aged 80 years or older (OR 1.17; 95% CI 0.97 to 1.42) (Figure 1).

Discussion

In this large, nationally representative U.S. cohort study, we found that COPD was an independent risk factor for inpatient mortality in patients hospitalized with COVID-19. This association was primarily driven by the effect in women and in adults between 40 and 79 years of age. This association persisted after adjustment for a wide range of potentially confounding variables, including age, race, medications received during hospitalization (remdesivir, systemic steroids), BMI, comorbidities, and use of mechanical ventilation.

Interestingly, all patients with COVID-19 and COPD had higher rates of ICU admission, mechanical ventilation, and palliative care consultation than those without COPD. Our results of worse outcomes amongst COVID-19 patients with COPD are consistent with previous research.^5-12

Our finding of worse mortality risk in women with COPD who were hospitalized due to COVID-19 is contrary to the overall outcomes of women with COVID-19 without COPD.^9,12,13,28 Why do women with COVID-19 who have COPD have increased risk of mortality? Prior to the COVID-19 pandemic, we knew that women with COPD have worse symptoms and airflow limitation compared to men, despite lower pack years of smoking.²⁹ Also, women’s COPD-related hospitalizations and deaths in the United States are worse compared to those of men.^30,31 In addition, women’s smaller lung size and other sex-specific factors may explain the observed association of worse mortality in women with COPD and COVID-19 compared to men.³² For example, estrogen has been reported to have an anti-inflammatory effect in pre-menopausal women^33,34 as well as against coronaviruses.³⁵ It is likely that the protective effect observed among women with COVID-19 has been driven primarily by pre-menopausal women.³³ Most women with COPD are post-menopausal²⁷ and likely have lost the protective effect of estrogen.^30-32

In addition to the interaction between sex and COPD, we found that patients with COPD who are 40 to 79 years old are at increased risk of death due to COVID-19, but those 80 years and older are not. Patients who are ≥80 years old have likely accumulated several comorbidities that may impair their health at baseline and ours (see Table 3) and other’s studies have consistently shown that older age and comorbidities are associated with worse outcomes from COVID-19.^{11,15,23,28,36} Very likely, these 2 variables have concealed any association between COPD and mortality from COVID-19 in this age group (≥80 years old). On the other hand, of particular concern is our finding of a 42% higher likelihood of death from COVID-19 in younger patients with COPD, those aged 40 to 64 years. These patients may have developed “early COPD” and possibly were exposed to significant smoking and other risk factors earlier in life that led to the development of likely anatomical, immunologic, and physiologic abnormalities³⁷ that may have made them more susceptible to die from COVID-19.

Although the interaction between COPD and ICS use on COVID-19 mortality in our cohort was non-significant, we did find that a history of an ICS prescription was associated with a lower risk of death in all hospitalized patients due to COVID-19. Despite this provocative finding, we do not know whether the ICS prescription had been filled. Use of an ICS was assumed if patients had a history of an ICS prescription 1 year prior to the COVID diagnosis or if the patient reported use. Furthermore, a recent United Kingdom-based observational study did not provide definitive answers to explain this association³⁸ and more studies would be necessary to further characterize this finding.

It is unclear what mechanisms drive the worse outcomes seen in patients with COPD hospitalized with COVID-19, but several biological factors have been proposed, including chronic lung inflammation, oxidative stress, protease-antiprotease imbalance, and increased airway mediators.^6,39,40 Additionally, patients with COPD have increased levels of angiotensin-converting enzyme 2, the receptor used by SARS-CoV-2 to enter host cells, that may enhance viral pathogenicity.^41-44 Moreover, viral infections can contribute to COPD exacerbations, leading to hospitalization, and acute exacerbations of COPD has been associated with poor outcomes⁴⁵

It is unknown whether patients with COPD have a higher risk of acquiring the SARS-CoV-2 infection. Studies based in the United Kingdom, Europe, and Asia report a varied prevalence of patients with COPD amongst hospitalized patients with COVID-19,^9,18,40 and COPD has been shown to be a risk factor for hospitalization due to COVID-19.¹⁵ The association between COPD, severe COVID-19, and mortality related to COVID-19 varies.^9-13,15,46 The current study represents the largest cohort studied to date of patients with COPD who have been hospitalized due to COVID-19 in the United States. Similarly, our findings of increased mortality associated with COPD support previous reports of poor outcomes in COVID-19 patients with chronic respiratory conditions. For example, patients with interstitial lung disease and COVID-19 were more likely to be hospitalized, require ICU care, and die compared to patients with COVID-19 without interstitial lung disease.^47,48

We acknowledge that this study has limitations, including its retrospective nature. The results merely show an association between COPD and mortality in hospitalized patients with COVID-19. Since we used ICD-10-CM codes to identify COPD, it is possible that our population may over- or under-represent the true prevalence of COPD in the general population.⁴⁹ In addition, we were not able to assess when remdesivir and systemic steroids were administered in relation to a patient’s hospitalization, the selection criteria used for drug prescription, or the duration of COVID-19 illness prior to the hospital admission. These unmeasured variables may have influenced whether the use of remdesivir and systemic corticosteroids affected clinical outcomes in hospitalized patients with COVID-19. Finally, patients with COPD have significant comorbid conditions,⁵⁰ and these may have mediated the association between COPD and increased hospital mortality. Yet, after adjustment for demographic and clinical factors, our findings indicated that COPD was an independent factor for mortality in patients hospitalized with COVID-19.

In conclusion, COPD is an independent risk factor for death in adults aged 40 to 79 years hospitalized due to COVID-19.

Acknowledgements

Author contributions: All authors have contributed to the current manuscript in the following manner:

• Substantial contribution to the conception, design, data acquisition, analysis and interpretation of the manuscript.
• Drafted and/or revised the manuscript for important intellectual content.
• Approved the final version to be submitted for publication.
• Agreed to be accountable for all aspects of the work and have ensured that questions related to the accuracy and integrity of any part of the manuscript are appropriately investigated and resolved.

The authors acknowledge and greatly appreciate the assistance in the preparation of this manuscript by: En Shuo Hsu, MA (Office of Biostatistics, University of Texas Medical Branch, Galveston, Texas); Sarah Toombs Smith, PhD, ELS (Research Communications Manager and Fellow, Sealy Center on Aging, University of Texas Medical Branch, Galveston, Texas); Tara N. Atkins, MLIS (Reference Librarian Moody Medical Library/Academic Resources, University of Texas Medical Branch, Galveston, Texas).

Declaration of Interest

Drs. Puebla Neira, Watts, Seashore, Duarte, Nishi, Baillargeon and Sharma, along with Ms. Polychronopoulou, have nothing to disclose. Dr. Kuo reports grants from the UTMB Claude D. Pepper Older Americans Independence Center and from the Agency of Healthcare Research and Quality during the conduct of the study.