Article Text
PDF
Abstract
Background COPD is a major cause of morbidity and mortality in populations eligible for lung cancer screening. We investigated the role of spirometry in a community-based lung cancer screening programme.
Methods Ever smokers, age 55–74, resident in three deprived areas of Manchester were invited to a ‘Lung Health Check’ (LHC) based in convenient community locations. Spirometry was incorporated into the LHCs alongside lung cancer risk estimation (Prostate, Lung, Colorectal and Ovarian Study Risk Prediction Model, 2012 version (PLCOM2012)), symptom assessment and smoking cessation advice. Those at high risk of lung cancer (PLCOM2012 ≥1.51%) were eligible for annual low-dose CT screening over two screening rounds. Airflow obstruction was defined as FEV1/FVC<0.7. Primary care databases were searched for any prior diagnosis of COPD.
Results 99.4% (n=2525) of LHC attendees successfully performed spirometry; mean age was 64.1±5.5, 51% were women, 35% were current smokers. 37.4% (n=944) had airflow obstruction of which 49.7% (n=469) had no previous diagnosis of COPD. 53.3% of those without a prior diagnosis were symptomatic (n=250/469). After multivariate analysis, the detection of airflow obstruction without a prior COPD diagnosis was associated with male sex (adjOR 1.84, 95% CI 1.37 to 2.47; p<0.0001), younger age (p=0.015), lower smoking duration (p<0.0001), fewer cigarettes per day (p=0.035), higher FEV1/FVC ratio (<0.0001) and being asymptomatic (adjOR 4.19, 95% CI 2.95 to 5.95; p<0.0001). The likelihood of screen detected lung cancer was significantly greater in those with evidence of airflow obstruction who had a previous diagnosis of COPD (adjOR 2.80, 95% CI 1.60 to 8.42; p=0.002).
Conclusions Incorporating spirometry into a community-based targeted lung cancer screening programme is feasible and identifies a significant number of individuals with airflow obstruction who do not have a prior diagnosis of COPD.
- lung cancer
- COPD epidemiology
Statistics from Altmetric.com
Key messages
What is the key question?
Is it feasible to incorporate spirometry into a community-based lung cancer screening programme in order to detect previously undiagnosed airflow obstruction?
What is the bottom line?
Incorporating spirometry into a community-based cancer screening programme, targeting high-risk populations, is feasible and identifies a significant number of individuals with airflow obstruction without a prior diagnosis of COPD.
Why read on?
Our study demonstrates how a ‘Lung Health Check’ approach to lung cancer screening, which includes spirometric assessment for undiagnosed COPD, can successfully reach individuals of low socioeconomic status from deprived areas and therefore provide an opportunity to address other smoking-related causes of morbidity and mortality in high-risk populations.
Introduction
Chronic Obstructive Pulmonary Disease (COPD) is a significant global health problem and a leading cause of morbidity and mortality with a considerable economic burden.1 2 The World Health Organization (WHO) has projected that deaths attributable to COPD will increase by up to 30% over the next 10 years and become the third leading cause of death across the globe.3 Tackling the prevalence of smoking, which is the most important risk factor for developing COPD, is a key goal.3 4 Individuals diagnosed with COPD often suffer from breathlessness, chronic cough and sputum production; however, during the early stages of the disease, it is often asymptomatic. COPD is therefore commonly underdiagnosed and also often misdiagnosed in primary care settings with reported prevalence rates ranging from 3.6% to 19%.5 6
Individuals with undiagnosed COPD are at increased risk of exacerbations and pneumonia; those with symptoms are also at increased risk of death.7 Understanding factors associated with delayed or missed diagnosis is therefore important.8 The burden on the healthcare system for the management of exacerbations is equivalent for individuals with airflow obstruction irrespective of previous COPD diagnosis.9 Diagnosis of COPD at an earlier stage may improve outcomes and quality of life4 although screening for asymptomatic COPD is not currently recommended.10 Characteristics reported to be associated with missed COPD diagnosis include: younger age, male gender, lower symptom burden, current smoking, lower education and having better lung function.6 7 For individuals with the relevant exposures and symptoms, spirometry is required to make the diagnosis. The key spirometric finding is airflow obstruction; this is defined as a postbronchodilator ratio of FEV1 to FVC of <0.7.4
Lung cancer is the world’s leading cause of cancer death and most patients are diagnosed at an advanced incurable stage primarily due to late presentation.11 Low-dose CT (LDCT) screening of high-risk smokers reduces lung cancer-specific mortality and a recently published position statement from European investigators recommended implementation of screening across Europe.12 The link between COPD and lung cancer is well established13 and the prevalence of COPD within lung cancer screening programmes well described.14–17 However, spirometry is not routinely incorporated into screening programmes. Including assessments for other smoking-related comorbidities such as COPD in screening programmes has the potential to improve cost effectiveness.18 Successful implementation of lung cancer screening will require services to be accessible to those at greatest risk, such as smokers with low socioeconomic status (SES).19 Low SES is associated with higher levels of smoking, COPD and lung cancer but also reduced participation in screening trials.20 21 We developed a ‘Lung Health Check’ programme which took place in deprived community locations across Manchester to try to address this participation bias and overcome practical barriers such as cost, distance and travel.22 The aim of this study was to determine whether incorporating spirometry into a community-based lung cancer screening programme was effective in identifying individuals with previously undetected airflow obstruction, and therefore possible COPD. We also wanted to explore the characteristics of individuals with airflow obstruction to see if there are important differences between those that do and those that do not have a prior diagnosis of COPD. We hypothesise that incorporating spirometry into a lung cancer screening programme would be feasible and would identify a significant number of attendees with previously undetected airflow obstruction.
Methods
Manchester Lung Health Checks
A detailed description of the Manchester Lung Health Check pilot has previously been published.23 24 In brief, ever smokers aged 55–74 registered with participating general practitioner (GP) practices were invited to attend an LHC in convenient community locations, across three sites, in deprived areas of Manchester. As part of the LHC, participants underwent an assessment of respiratory symptoms, spirometry and a calculation of their 6-year lung cancer risk score using the Prostate, Lung, Colorectal and Ovarian Study Risk Prediction Model, 2012 version (PLCOm2012).25 This model takes several parameters into account including age, gender, ethnicity, smoking exposure, education, body mass index, previous medical and family history. Those at a higher risk of lung cancer, defined as a PLCOm2012 score of ≥1.51%,26 were invited to participate in annual LDCT screening over two screening rounds. Postcodes were recorded to determine Index of Multiple Deprivation 2015 (IMD) rank for England. This is a measure of relative deprivation in small areas (neighbourhoods) of England; areas are ranked from 1 (most deprived) to 32 844 (least deprived).27 Healthcare records were searched for a diagnosis of lung cancer using International Classification of Diseases (ICD) codes in all those lost to follow-up in the second round of screening.
Spirometry
Spirometry was performed in accordance with Association for Respiratory Technology & Physiology (ARTP) guidelines28 at the time of the LHC using a desktop spirometer (Vitalograph ALPHA). All nurses performing spirometry had undergone formal ARTP training. All measures were prebronchodilator and included FEV1, FVC and their respective Global Lung Function Initiative reference values. Each participant was given at least three attempts at performing spirometry and the highest value recorded. Airflow obstruction was defined as an FEV1/FVC ratio <0.7.4
Symptom questionnaire
Participants were asked about respiratory symptoms at the time of the LHC including: breathlessness (using the Medical Research Council (MRC) dyspnoea scale), cough (including duration), sputum production, wheeze, haemoptysis, chest pain and antibiotic or steroid use for exacerbations (≥2 episodes per year). In accordance with Global Initiative for Chronic Obstructive Lung Disease recommendations, participants were categorised as ‘symptomatic’ if they reported any one or more of the following symptoms: breathlessness (MRC dyspnoea scale ≥2), chronic cough (≥6 weeks) or regular sputum production (≥teaspoon/day).4
Definition of COPD
For the purpose of these analysis, individuals were categorised as potentially having COPD if they had airflow obstruction, independent of presence and pattern of symptoms. Participants were asked if they had ever been diagnosed with COPD or emphysema and primary care records were searched for a prior diagnosis of COPD, emphysema or chronic bronchitis using ICD codes in all those with airflow obstruction; those without a prior diagnosis were categorised as ‘undiagnosed’.
LDCT and radiological reporting
All LDCT scans (Optima 660, GE Healthcare) were reported by National Health Service Consultant Radiologists with a specific interest in thoracic radiology. As part of the reporting template, other respiratory radiological findings were recorded including the presence or absence of emphysema. Such findings were reported at the discretion of the reporting radiologist without predefined definitions. The severity or form of emphysema was not recorded.
Statistical analysis
All statistical analysis was conducted using SPSS V.22. Groups were compared using independent t-test (parametric) or Mann-Whitney U test (non-parametric) for continuous data and χ2 and Fisher’s exact tests for categorical data. We tested whether the presence of airflow obstruction was independently predictive of lung cancer after 2 years of follow-up by undertaking a logistic regression analysis including PLCOM2012 score (as this model includes a number of variables associated with lung cancer risk). In further analyses limited to those with airflow obstruction, we performed univariate binary logistic regression to determine whether standard clinical variables differed between individuals with airflow obstruction who did or did not have a prior COPD diagnosis. Variables with a significant association (p≤0.05) were then entered into a multivariate model. We also tested whether a GP diagnosis of COPD was independently predictive of lung cancer by undertaking a logistic regression analysis including PLCOM2012 score. Statistical significance was defined as p≤0.05.
Results
Spirometry
Of 2541 individuals who attended an LHC, 99.4% successfully performed spirometry (n=2525/2541). Of these 37.4% (n=944) had evidence of airflow obstruction. A comparison of baseline characteristics between those with and without airflow obstruction is detailed in table 1. Those with airflow obstruction had greater tobacco smoke exposure (smoking duration 40±13 vs 31±15 years; p<0.001), were more likely to be current smokers (49% vs 27%; p<0.001), were more symptomatic (68% vs 55%; p<0.001) and had a lower median IMD deprivation rank (2806 vs 2908; p=0.004). Mean predicted FEV1 was 99%±21% (2.6±0.7 L) in those without and 76%±24% (1.9±0.7 L) in those with airflow obstruction and the corresponding FEV1/FVC ratio was 0.77±0.05 and 0.60±0.09, respectively (p<0.001).
Baseline characteristics of all participants stratified according to previous COPD diagnosis in those with obstructive spirometry
Participants were asked, as part of the LHC, to self-report whether they had ever been diagnosed with COPD as this is a variable used in the PLCOM2012 model to calculate lung cancer risk. Twelve per cent (n=193) of those without airflow obstruction reported a previous diagnosis of COPD and 61% (n=576/944) with airflow obstruction reported no prior diagnosis of COPD. If we assume a direct relationship between airflow obstruction and COPD, then up to 30.5% (n=769/2525) of LHC attendees may have been misclassified. This varied from 22.7% in those not eligible for screening (n=251/1108) to 36.6% in those who were eligible (n=518/1417).
Previous diagnosis of COPD
A primary care medical case record review was undertaken in all those with airflow obstruction (n=944) to look for evidence of a prior diagnosis of COPD. In just over half of individuals, a diagnosis of COPD was documented (50.3%; n=475; figure 1). All individuals with obstructive spirometry were divided into two groups, those with a documented history of COPD and those without in order to identify characteristics associated with a missed COPD diagnosis. The presence of airflow obstruction without a prior COPD diagnosis was associated, after multivariate analysis, with younger age (p=0.015), male sex (adjOR 1.84, 95% CI 1.37 to 2.47; p<0.0001), lower smoking duration (p<0.0001), mean cigarettes per day (p=0.035), being asymptomatic (adjOR 4.19, 95% CI 2.95 to 5.95; p<0.0001) and higher FEV1/FVC ratio (p<0.0001; table 2).
Consolidated Standards of Reporting Trials diagram showing flow of participants through study.
Univariate and multivariate analysis of the association between various clinical measures and the presence of airflow obstruction without a prior diagnosis of COPD
CT finding of emphysema
1374(54.4%) of those who performed spirometry were eligible for and underwent LDCT screening. The prevalence of radiological emphysema was high, with emphysema detected in 63.4% (n=876) of individuals; this increased to 76.0% (n=529/696) in those with airflow obstruction compared with 51.2% (n=347/678) in those without (p<0.001). The detection of emphysema was higher in those with airflow obstruction and a previous diagnosis of COPD compared with those with airflow obstruction and no previous diagnosis of COPD (83.9% vs 65.4%; p<0.001).
Screen detected lung cancers
In those with evidence of airflow obstruction, participants previously diagnosed with COPD had a higher mean lung cancer risk score than those without (mean PLCOm2012 score 6.4±5.5% vs 3.0±3.1%; p<0.001; table 1). A total of 61 lung cancers were detected in the screened population. Within the 2 years of follow-up, 12 participants had died and 61 (4.9%) did not attend their second round scan, none of whom had a diagnosis of lung cancer on ICD code searches. Lung cancer detection was significantly higher in those with airflow obstruction (5.7% vs 3.1%; p<0.001; table 1). However, airflow obstruction was not independently predictive of lung cancer, when adjusted for PLCOM2012 lung cancer risk score (p=0.29).
Among participants with airflow obstruction, lung cancer detection was also significantly higher in those with a COPD diagnosis compared with those with previously unrecognised airflow obstruction (6.9% vs 1.5%; p<0.001; table 1). The risk of lung cancer was significantly higher in those with diagnosed COPD (OR 3.76, 95% CI 1.64 to 8.62; p=0.002) and this remained significant after adjustment for PLCOM2012 risk score (adjOR 2.80, 95% CI 1.60 to 8.42; p=0.002).
Discussion
In this study, we investigated the role of spirometry in a community-based lung cancer screening programme. Airflow obstruction, a key requirement for the diagnosis of COPD, was found in more than a third of participants (37.4%). Half of these individuals had no previous diagnosis of COPD, equivalent to one in five (18.6%) of all LHC attendees. Although those without a prior diagnosis of COPD were less symptomatic than those known to have COPD, the prevalence of symptoms was still high, present in more than half (53.3%). As a consequence, 1 in 10 (9.9%) screening attendees were symptomatic with previously unrecognised airflow obstruction, raising the possibility of COPD. A further 8.7% were asymptomatic but had airflow obstruction. Both groups have increased morbidity and, in the case of symptomatic undiagnosed COPD, increased mortality.7 In those with evidence of obstructive spirometry, individuals were less likely to be diagnosed with COPD if they were male, younger, had less tobacco smoke exposure, were asymptomatic and had better lung function.
Our study demonstrates a high prevalence of undetected airflow obstruction among lung cancer screening participants. This is in keeping with previous studies14–16; however, populations recruited to large research trials are generally more educated and of higher SES than would be expected in those at high risk of lung cancer more broadly.20 21 Attendees of Manchester’s LHCs were from highly deprived areas and may therefore represent a more ‘real-world’ population not previously captured. We believe that this makes our results highly pertinent to screening implementation. The findings emphasise that COPD and lung cancer coexist and that future lung cancer risk prediction would benefit from more precise COPD diagnosis in primary care. Our study also adds to the pre-existing literature which has examined factors associated with undiagnosed COPD. Consistent with previous large studies, we demonstrate being male, less symptomatic and with better lung function to be associated with having undiagnosed COPD.7 8 However, in contrast to Çolak et al 7 and Martinez et al, 8 we report being a former smoker to be associated with undiagnosed COPD as well. Case finding programmes targeting current and ex-smokers have had only limited impact29 30 and therefore incorporating COPD diagnosis into an LHC-based lung cancer screening programme may improve reach to target populations. Interestingly, we report an almost threefold increased risk of screen detected lung cancer in those with airflow obstruction previously diagnosed with COPD compared with those without a COPD diagnosis. The exact reasons for this are unclear but could possibly be related to the higher levels of smoking exposure, higher degrees of airflow obstruction and higher prevalence of radiological emphysema in this group, all of which have been shown to have an effect on the significance of the impact of COPD on lung cancer risk.31
Emphysema was highly prevalent in our screened cohort with evidence of emphysema in two-thirds (63.4%) of individuals. The presence of emphysema was not associated with screen detected lung cancer but was more common in individuals with airflow obstruction (76%). Individuals with obstructive spirometry and emphysema on CT were 2.5 times more likely to have had a previous diagnosis of COPD than those without emphysema.
The Manchester screening pilot is one of the first programmes to use an LHC approach to identify high-risk individuals for LDCT lung cancer screening. Participants of the programme had a median deprivation rank within the lowest decile for England suggesting successful engagement with high-risk individuals of lower SES. The incorporation of spirometry into the LHC was straightforward and identified a large number of people with previously undetected airflow obstruction. We acknowledge that results are based on prebronchodilator spirometry readings at a single timepoint which is not optimal for COPD assessment as this may include individuals with reversible airflow obstruction (eg, asthma). Our results may therefore overestimate the number of undiagnosed COPD cases. We did not label participants as having COPD but communicated the results to primary care with a recommendation for more detailed assessment.
Unlike many previous studies that rely on self-reported COPD diagnosis and are therefore subject to recall bias, the use of GP medical records allowed for a more accurate recording of previous COPD diagnosis, allowing for the fact primary care records also have inaccuracies and limitations. This is important as with all survey-based studies there is a reliance on self-reported symptoms, smoking exposure and medical history which may have inherent inaccuracies. In our cohort, for almost one in three attendees, there was discordance between the self-reported presence of COPD and spirometry results. Such inaccuracies can have implications for lung cancer risk prediction models, such as the PLCOm2012 used in our study, as participants could be misclassified as high or low risk.
In conclusion, our study demonstrates that a community-based targeted lung cancer screening programme can successfully incorporate spirometry and identify a significant number of individuals with previously unrecognised airflow obstruction, both symptomatic and asymptomatic. Our LHC approach was successful in reaching individuals with low SES from deprived areas and may therefore provide an opportunity to address a significant cause of morbidity and mortality in high-risk populations. Currently, there is a lack of evidence to support screening for COPD in asymptomatic individuals.10 However, early diagnosis can allow an opportunity for a number of beneficial interventions such as smoking cessation, education, vaccination, exercise and rehabilitation, aggressive management of comorbidities as well as pharmacotherapy which could lead to better outcomes and improved quality of life in these individuals.4 In addition to the health benefits associated with this approach, there is also the potential to improve cost effectiveness of such a screening programme especially if combined with other targeted initiatives such as screening for cardiovascular disease32 and smoking cessation.33 The true benefits of incorporating COPD diagnosis into lung cancer screening services, especially in asymptomatic individuals, are therefore an important question for future programmes to address.
Acknowledgments
JV and PAC are supported by the NIHR Manchester Biomedical Research Centre.
References
Footnotes
Twitter @hsbalata, @JetstreamSol, @DrPhilCrosbie
Contributors Service concept: RB, PAC, PVB and JT. Service development by members of the Macmillan Cancer Improvement Partnership: RB, PAC, PVB, JT, DC, PE and ST. Service operation and delivery by the Manchester University NHS Foundation Trust lung cancer team: HB, ME, RB, PAC and AS. Radiology reporting by the radiology consortium: RD, MG, JH, KI, DK, SM, TN, RS, AS, ES, BT, AW and JW. Data collection, analysis and drafting of manuscript: HB, JH, JV, RB and PAC. Guarantors of overall content: HB, RB and PAC. Review, revision and agreement of final manuscript: all authors.
Funding This study was funded by Macmillan Cancer Improvement Partnership.
Competing interests None declared.
Patient consent for publication Not required.
Ethics approval Clinical data from the screening service were recorded on an ethically approved database (REC Ref: 16/NW/0013). North West-Greater Manchester West Research Ethics Committee.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement Data are available upon reasonable request.
Linked Articles
- Airwaves