Association between pre-existing respiratory disease and its treatment, and severe COVID-19: a population cohort study

Paul Aveyard, Min Gao, Nicola Lindson, Jamie Hartmann‐Boyce, Peter Watkinson, Duncan Young, Carol Coupland, Pui San Tan, Ashley Kieran Clift, David A Harrison, Doug W Gould, Ian Pavord, Julia Hippisley–CoxThe Lancet Respiratory Medicine 2021 · DOI

TL;DR

This large-scale population cohort study investigated the association between pre-existing respiratory conditions and severe COVID-19 outcomes in England. It found that while conditions like COPD and interstitial lung disease are associated with a modestly increased risk of severe COVID-19, the risk for asthma is relatively small, and the overall risk of death from COVID-19 for individuals with these conditions was often lower than their ordinary risk of death from any cause.

What Problem Does This Paper Address?

Early in the COVID-19 pandemic, there was a prevailing concern that individuals with pre-existing respiratory diseases, such as asthma and COPD, would be at a significantly higher risk of severe illness, hospitalization, intensive care unit (ICU) admission, and death from COVID-19. This concern was amplified by anecdotal evidence and some early case series suggesting a lower prevalence of chronic respiratory disease in hospitalized COVID-19 patients, leading to questions about whether these conditions might offer some protection or if the data was being misinterpreted. The aim of this study was to provide robust, population-level evidence to clarify these associations, moving beyond hospital-based cohorts which can be prone to bias, and to investigate the role of inhaled corticosteroid (ICS) treatments.

Methods

The researchers employed a population cohort study design, leveraging a large dataset from 1205 general practices in England contributing to the QResearch database. This database was linked to Public Health England's SARS-CoV-2 testing records, English hospital admission data (Hospital Episode Statistics), ICU admission data (Intensive Care National Audit & Research Centre Case Mix Programme), and death certificates (Office for National Statistics).

The primary analytical method used was Cox's proportional hazards models to estimate hazard ratios (HRs) for the association between various respiratory diseases and severe COVID-19 outcomes (hospitalization, ICU admission, and death). These models were meticulously adjusted for a comprehensive set of potential confounders, including demographic factors (age, sex, ethnicity, socioeconomic status, region), body-mass index (BMI), smoking status, and a wide range of comorbidities (both smoking-related and non-smoking-related illnesses).

A post-hoc analysis specifically examined the role of inhaled corticosteroids (ICS) by classifying individuals based on their prescription history. The study also conducted sensitivity analyses, including excluding care home residents to account for their unique vulnerability, and examined potential effect modification by age, sex, ethnicity, and smoking status.

Datasets

The study utilized several large, established datasets:

  • QResearch database: A collection of anonymized primary care records from 1205 general practices in England, covering patient consultations, diagnoses, and prescriptions.
  • Public Health England's SARS-CoV-2 testing database: Provided data on COVID-19 test results.
  • Hospital Episode Statistics (HES): Contained data on hospital admissions in England.
  • Intensive Care National Audit & Research Centre (ICNARC) Case Mix Programme database: Provided data on ICU admissions.
  • Office for National Statistics (ONS) death certificates: Supplied data on mortality.

No new benchmarks were introduced; the study focused on analyzing existing, comprehensive health records.

Key Insights and Ideas

  • Asthma's modest risk: Contrary to some early concerns, asthma was associated with only a small increase in the risk of severe COVID-19 outcomes, particularly hospitalization. The risk of death was not significantly increased.
  • COPD and interstitial lung disease carry a higher, but still modest, risk: Conditions like COPD and interstitial lung diseases were linked to a more substantial, though still modest, increase in the risk of severe COVID-19 compared to the general population.
  • Risk of death often lower than ordinary risk: For many respiratory conditions, the risk of death from COVID-19, even at the height of the epidemic, was often far lower than the ordinary risk of death from any cause for individuals in those age groups. This suggests that pre-existing respiratory disease, while a risk factor, might not be as dire a predictor of mortality as initially feared.
  • ICS use and risk remain unclear: The study found that while ICS use might be associated with a modestly increased risk of severe COVID-19, this association was reduced after accounting for other risk factors. The overall implication is that ICS use does not appear to significantly influence the likelihood of severe COVID-19, and its role remains complex.
  • Age is a critical factor: The association between respiratory disease and severe COVID-19 was significantly modified by age, with younger individuals showing a more pronounced increase in risk for hospitalization and ICU admission, potentially linked to higher BMI in this age group.
  • Gender differences: Women with respiratory diseases showed a statistically significantly higher relative hazard of hospitalization, ICU admission, and death compared to men, except in one specific analysis.

Key Results

  • Hospitalization:
    • COPD: HR 1.54 (95% CI 1.45-1.63)
    • Asthma: HR 1.18 (95% CI 1.13-1.24)
    • Severe asthma: HR 1.29 (95% CI 1.22-1.37)
    • Interstitial lung diseases: HRs ranged from 1.59 to 1.66.
  • Lung cancer: HR 2.24 (95% CI 1.89-2.65)
  • ICU Admission:
    • Asthma: HR 1.08 (95% CI 0.93-1.25)
    • Severe asthma: HR 1.30 (95% CI 1.08-1.58)
    • COPD: HR 0.85 (95% CI 0.65-1.11) (no increased risk)
  • Death:
    • COPD: HR 1.54 (95% CI 1.42-1.67)
    • Asthma: HR 0.99 (95% CI 0.91-1.07) (no increased risk)
    • Interstitial lung diseases: HRs ranged from 1.47 to 2.05.
  • Lung cancer: HR 1.77 (95% CI 1.37-2.29)
  • ICS Use: Associated with a modest increased risk of severe COVID-19 (HR 1.13 for hospitalization, 1.63 for ICU, 1.15 for death), but this was reduced when other risk factors were considered.

Limitations and Future Work

  • Severity of disease not captured: The study classified respiratory diseases but did not capture their severity. This means the risks might be overestimated for individuals with mild disease and underestimated for those with more severe conditions.
  • ICS classification: The definition of regular ICS use (at least two prescriptions) might have misclassified some users and non-users, potentially moving estimated associations towards the null.
  • ICU care as a proxy for severity: Using ICU admission as a sole indicator of severe COVID-19 is a limitation, as many frail individuals may not be admitted to ICU even if critically ill.
  • "Long COVID" not assessed: The study's dataset did not allow for the assessment of long-term COVID-19 symptoms.
  • Potential for underestimation of risk: It's possible that individuals with respiratory diseases, aware of their increased risk, may have adopted more rigorous behavioral measures (e.g., social distancing), leading to an underestimation of their true risk.
  • Future work: The authors suggest that advice to individuals with respiratory disease should be more reassuring, given the findings. Further research could explore the specific impact of different severities of respiratory disease and the long-term consequences of COVID-19 in these populations. The study also highlights the need for more nuanced understanding of the interplay between respiratory conditions, treatments, and COVID-19 outcomes.

Summary updated

Abstract

National Institute for Health Research Oxford Biomedical Research Centre and the Wellcome Trust.
BibTeX
@inproceedings{aveyard2021association,
  author = {Paul Aveyard and Min Gao and Nicola Lindson and Jamie Hartmann‐Boyce and Peter Watkinson and Duncan Young and Carol Coupland and Pui San Tan and Ashley Kieran Clift and David A Harrison and Doug W Gould and Ian Pavord and Julia Hippisley–Cox},
  title = {{Association between pre-existing respiratory disease and its treatment, and severe COVID-19: a population cohort study}},
  booktitle = {The Lancet Respiratory Medicine},
  year = {2021},
  doi = {10.1016/s2213-2600(21)00095-3},
}
Figure