Background
The COVID-19 pandemic and subsequent challenges to healthcare systems have led to unprecedented disruptions of routine care for people with chronic conditions. In response to a surge in COVID-19 cases, the UK and devolved governments announced the first nationwide lockdown on 23rd March 2020, thereby severely restricting movements and social contacts [
1]. The accompanying messages to avoid overwhelming the National Health Service (NHS) and fear of contracting SARS-CoV-2 in healthcare settings had an impact on people’s willingness to seek emergency care [
2].
There is evidence that the UK-wide lockdown was associated with poorer cardiovascular [
3] and cancer [
4] outcomes, but its impact on serious chronic obstructive pulmonary disease (COPD) outcomes remains unclear.
Despite the disruption to routine COPD care [
5,
6], the societal changes associated with lockdown—in particular, improvements in air quality and reductions in other viruses responsible for acute respiratory tract infections [
7,
8]—may have led to an overall improvement in COPD outcomes. The available body of evidence suggests that there may have been a reduction in acute exacerbations of COPD (AECOPD), but these data are difficult to interpret because of methodological limitations including studying selective populations and/or from a limited number of centres [
9‐
26]. Despite the changes in healthcare-seeking behaviour during the lockdown, people with severe AECOPD were still likely to seek medical attention as the symptoms are intense such that they are difficult to tolerate at home [
12].
We sought to investigate the impact of the UK-wide COVID-19 lockdown on the overall numbers of recorded severe AECOPDs leading to admission and/or death across the entire populations of Scotland and Wales. To contextualise the findings, we also investigate AECOPDs that were recorded in primary care and emergency departments (EDs) in Wales.
Discussion
Our national level interrupted time-series analyses of the impact of lockdown across Scotland and Wales found substantial reductions in severe AECOPD leading to ED attendance and/or hospital admission as well as the less severe AECOPD that are recorded in primary care. The levels remained well below the corresponding 5-year averages throughout the study period. The easing of strict lockdown restrictions announced at the end of May 2020 [
38,
39] did not lead to a substantial rebound in these events, but there was a gradual rise in ED attendances and admissions. There was no evidence of an increase in deaths due to COPD during the lockdown.
These findings are especially significant given that COPD exacerbations are one of the commonest reasons for emergency admission to hospital [
40]. Reduced incidence of AECOPDs therefore increases healthcare capacity and resources available for those with COVID-19, as well as conferring obvious benefits to people with COPD. A reduction in AECOPD admissions is also particularly advantageous given the specific challenges in the hospital management of AECOPD during the pandemic—components of care, such as non-invasive ventilation (NIV), are associated with increased risk of viral transmission due to aerosolization [
41] and can therefore only be delivered in specified clinical areas.
To the best of our knowledge, this is the first national level analysis of the impact of the COVID-19 lockdown on AECOPD incidence and mortality, using data across primary and secondary care as well as data on deaths. We used population-based data with high-to-complete geographical coverage across Scotland and Wales, which enabled comparison of findings between the two UK nations, which were broadly comparable.
Our study has some limitations. Firstly, there are no validated case definitions for COPD admissions or deaths in UK data, and so the case definitions used may have variable accuracy. We did not include deaths with COPD as a contributing cause because they would not have been specific enough for the purpose of this study. Data on AECOPD in primary care and ED data is under-recorded [
42,
43]. However, our sensitivity analysis, limiting data to those aged ≥ 35 and ever-smokers, did not impact our overall results. Further, given that we were interested in trends over time rather than absolute numbers and that coding practices of these events are unlikely to have changed, this is unlikely to have significantly affected our findings.
The observed reduction in AECOPDs during lockdown does not necessarily imply direct causal effects of lockdown. Nonetheless, they are likely to have been mediated by reductions in the transmission of other respiratory pathogens and outdoor air pollution during lockdown [
7,
8,
14], both of which have a major role in triggering AECOPD [
44]. These factors could also have mediated the reductions in emergency admissions for asthma that have been reported following lockdowns [
45‐
48].
A number of potential confounding factors such as changes in prescribing for COPD [
49], behavioural changes related to improved self-management and smoking reduction/cessation [
50], and possibly improved air quality [
7] during the first wave of pandemic might have contributed to the observed reduction in AECOPDs. However, we were not in a position to adjust for these potential sources of bias. Care therefore needs to be taken when interpreting our findings. Furthermore, national messaging on the need to avoid overwhelming the NHS and the fear about the spread of SARS-CoV-2 might have contributed to the fall in AECOPD-related ED attendances and admissions during the first wave of the pandemic. However, this effect on healthcare-seeking behaviour is likely limited in those experiencing a severe AECOPD, which usually requires hospital assessment and treatment. In addition, the corresponding decline in AECOPD in primary care records suggests a true decline in incidence during lockdown. However, it is possible that the milder forms of AECOPDs have been self-managed by patients and were not presented to the health care system during lockdown.
The lifting of restrictions on travel and social contact was associated with a gradual rise in emergency admissions for AECOPD seen in our data for Scotland and Wales. This could be partly due to an increase in the circulation of respiratory viruses (mainly rhinoviruses initially) [
8] and levels of outdoor air pollutants [
51]. However, this could also be explained by an increasing threshold of hospital admission, since primary care consultations for AECOPD continued to fall towards summer in accordance with the typical seasonal trend of AECOPD in the UK [
52].
Although there is extensive literature on COPD as a risk factor for COVID-19 severe outcomes and deaths [
53‐
55], there is currently limited data examining how the pandemic, and specifically lockdowns, has affected COPD deaths more widely. Two studies from Hong Kong and England reported no difference in inpatient mortality during admissions for COPD exacerbation [
13,
15]. Analysis of excess mortality from specific conditions is important in understanding whether reductions in emergency healthcare utilisation represent a true reduction in incidence or avoidance of healthcare settings, the latter of which could lead to increased mortality. Our study has shown no significant increase in non-COVID COPD deaths over the first 30 weeks of 2020. This is in contrast to a recent analysis of cardiovascular mortality in England and Wales [
3], which demonstrated excess non-COVID acute cardiovascular deaths. The authors suggested that people with these diagnoses either did not seek help for their illness or were not referred to hospital, consistent with the greatest proportional increase in cardiovascular deaths occurring in community settings. Data from England from March to September 2020 from the Office for National Statistics shows that COVID-19 accounted for over 90% of excess deaths among those aged over 75 from both sexes, but the proportion of non-COVID excess deaths was higher across younger people [
56]. The leading causes of these non-COVID excess deaths were dementia, ischaemic heart disease, cerebrovascular disease and other circulatory diseases. Deaths in England due to “chronic lower respiratory diseases” (ICD-10 codes J40-47) including COPD actually fell compared to expected levels when assessed cumulatively from March to September 2020 [
56], consistent with our findings in Scotland and Wales.
There are several important areas for future investigation to understand underlying reasons for our findings. These include person-level analyses of how factors related to COPD, such as disease severity, control and health service utilisation and positive drivers such as reduced exposure to respiratory pathogens and pollutants, improved self-management, smoking cessation and other behavioural changes, might have affected the risk of AECOPD and related death during lockdown. If further work suggests that altered outdoor air pollution levels have played a significant role, findings should spur increased drive to improve air quality longer term [
57]. Other interventions with the potential to produce lasting reductions in the rate of severe AECOPD include the facilitation of self-management of chronic conditions and the consolidation of public health messages to reduce the transmission of respiratory infections, including hand hygiene, use of facemasks and wider deployment of testing and isolation when viruses are most likely to be circulating.
Acknowledgements
This study made use of anonymised data held at PHS and the SAIL Databank. We would like to acknowledge all the data providers who make anonymised data available for research.
EAVE II Collaborators
Colin R Simpson, Wellington School of Health, Faculty of Health, Victoria University of Wellington, Wellington, New Zealand and Usher Institute, The University of Edinburgh, Edinburgh, UK
Jim McMenamin, Public Health Scotland, Glasgow, UK
Lewis D Ritchie, Centre of Academic Primary Care, University of Aberdeen, Aberdeen, UK
Mark Woolhouse, Usher Institute, The University of Edinburgh, Edinburgh, UK
Helen R Stagg, Usher Institute, The University of Edinburgh, Edinburgh, UK
Diogo Marques, Public Health Scotland, Glasgow, UK
Josie Murray, Public Health Scotland, Glasgow, UK
Sarah Stock, Usher Institute, The University of Edinburgh, Edinburgh, UK
Rachael Wood, Public Health Scotland and University of Edinburgh, UK
Colin McCowan, School of Medicine, University of St Andrews, St Andrews, UK
Utkarsh Agrawal, School of Medicine, University of St Andrews, St Andrews, UK
Annemarie B. Docherty, Usher Institute, The University of Edinburgh, Edinburgh, UK
Rachel H. Mulholland, Usher Institute, The University of Edinburgh, Edinburgh, UK
Emily Moore, Public Health Scotland, Glasgow, UK
James Marple, Royal Infirmary of Edinburgh, Edinburgh, UK
Vicky Hammersley, Usher Institute, The University of Edinburgh, Edinburgh, UK
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.