Skip to main content
NCBI home page
JAMA Network logoLink to JAMA Network
. 2025 Jan 22;333(9):784–792. doi: 10.1001/jama.2024.24212

Polygenic Risk Score Added to Conventional Case Finding to Identify Undiagnosed Chronic Obstructive Pulmonary Disease

Jingzhou Zhang 1,2,3, Brian D Hobbs 3,4, Edwin K Silverman 3,4, David Sparrow 1,2,5, Victor E Ortega 6, Hanfei Xu 7, Chengyue Zhang 3, Josée Dupuis 7,8, Allan J Walkey 9, George T O’Connor 1,2, Michael H Cho 3,4, Matthew Moll 3,4,10,✉
PMCID: PMC11880956  PMID: 39841442

Key Points

Question

Does a polygenic risk score, which summarizes a person’s genetic risk for chronic obstructive pulmonary disease (COPD), enhance the identification of undiagnosed COPD cases beyond a case-finding questionnaire based on conventional risk factors and respiratory symptoms?

Findings

In this cross-sectional analysis of adults who reported no history of physician-diagnosed COPD from the Framingham Heart Study of community-dwelling adults and the Genetic Epidemiology of COPD (COPDGene) study of individuals who previously or currently smoke, adding a COPD polygenic risk score to the modified Lung Function Questionnaire score significantly improved discrimination performance in both cohorts and classification of COPD cases in the Framingham Heart Study.

Meaning

A COPD polygenic risk score may provide additional clinical utility in identifying undiagnosed COPD cases beyond conventional case-finding approaches, particularly in the general population.

Abstract

Importance

Chronic obstructive pulmonary disease (COPD) is often undiagnosed. Although genetic risk plays a significant role in COPD susceptibility, its utility in guiding spirometry testing and identifying undiagnosed cases is unclear.

Objective

To determine whether a COPD polygenic risk score (PRS) enhances the identification of undiagnosed COPD beyond a case-finding questionnaire (eg, the Lung Function Questionnaire) using conventional risk factors and respiratory symptoms.

Design, Setting, and Participants

This cross-sectional analysis of participants 35 years or older who reported no history of physician-diagnosed COPD was conducted using data from 2 observational studies: the community-based Framingham Heart Study (FHS) and the COPD-enriched Genetic Epidemiology of COPD (COPDGene) study.

Exposures

Modified Lung Function Questionnaire (mLFQ) scores and COPD PRS.

Main Outcomes and Measures

The primary outcome was spirometry-defined moderate to severe COPD (forced expiratory volume in the first second of expiration/forced vital capacity [FEV1/FVC] <0.7 and FEV1 [percent predicted] <80%). The performance of logistic models was assessed using the PRS, mLFQ score, and PRS plus mLFQ score for predicting spirometry-defined COPD.

Results

Among 3385 FHS participants (median age, 52.0 years; 45.9% male) and 4095 COPDGene participants (median age, 56.8 years; 55.5% male) who reported no history of COPD, 160 (4.7%) FHS and 775 (18.9%) COPDGene participants had spirometry-defined COPD. Adding the PRS to the mLFQ score significantly improved the area under the curve from 0.78 to 0.84 (P < .001) in FHS, 0.69 to 0.72 (P = .04) in COPDGene non-Hispanic African American, and 0.75 to 0.78 (P < .001) in COPDGene non-Hispanic White participants. At a risk threshold for spirometry referral of 10%, the addition of the PRS to the mLFQ score correctly reclassified 13.8% (95% CI, 6.6%-21.0%) of COPD cases in FHS, but not in COPDGene.

Conclusions and Relevance

A COPD PRS enhances the identification of undiagnosed COPD beyond a conventional case-finding approach in the general population. Further research is needed to assess its impact on COPD diagnosis and outcomes.

This cross-sectional study examines whether a COPD polygenic risk score enhances the identification of undiagnosed COPD beyond a conventional case-finding approach in the general population.

Introduction

Chronic obstructive pulmonary disease (COPD) is highly prevalent and deadly, with an increasing projected global disease burden in the coming decades.1 However, up to 70% of people with obstructive lung diseases remain undiagnosed in the US and worldwide.2,3 Compared with healthy controls, individuals living with undiagnosed COPD have reduced quality of life, work productivity, and survival.3,4,5,6 Although people with undiagnosed COPD typically have milder symptoms and less severe airflow obstruction than those with diagnosed COPD, a substantial proportion of individuals with overt symptoms and moderate to severe COPD remain undiagnosed.7 These individuals may miss opportunities for effective treatment that can alleviate symptoms and reduce exacerbations. Additionally, preventive interventions, such as smoking cessation, are crucial for all individuals with COPD, regardless of disease severity.

Case finding for COPD, aiming to identify at-risk individuals and refer them for confirmatory spirometry testing, is a cost-effective approach to diagnosing previously unrecognized cases.8,9,10 Contemporary case-finding tools can include questionnaires that assess risk factors and respiratory symptoms, sometimes combined with handheld devices that assess peak expiratory flow.11,12 Identifying patients with undiagnosed COPD may improve their health outcomes. A recent clinical trial demonstrated that active community-based case findings, coupled with pulmonologist-directed treatment, reduced health care use for respiratory illnesses by approximately half in individuals with previously unrecognized COPD compared with controls.13

The heritability of COPD is estimated to be approximately 40% in unrelated individuals, indicating that genetic factors contribute substantially to the development of COPD.14 Although using genetic risk factors in COPD case-finding strategies has been suggested,12 it remains unknown whether incorporating genetic risk measures into contemporary case-finding tools can improve the identification of undiagnosed COPD. For complex traits such as COPD, genetic risk is largely determined by many common genetic variants, each contributing a small effect. Aggregating the effects of numerous genetic variants across the genome into a polygenic risk score (PRS) can improve predictive power for the disease.15

A COPD PRS that was associated with both prevalent COPD and age of COPD diagnosis was previously developed.16,17 However, the performance of the PRS within a case-finding framework and individual-level predictive metrics, such as calibration plots, that provide deeper insights into its predictive accuracy were not assessed. It was hypothesized that the PRS would improve the identification of individuals with undiagnosed moderate to severe COPD when combined with the conventional case-finding approach. To test this hypothesis, this study examined the role of the PRS in community-dwelling adults from the Framingham Heart Study (FHS) and COPD-enriched people who reported ever smoking from the Genetic Epidemiology of COPD (COPDGene) study.

Methods

Study Populations

We included participants 35 years or older without a self-reported history of physician-diagnosed COPD from the FHS and COPDGene studies, both of which have been previously described.18,19,20 Briefly, FHS recruited community-based individuals predominantly of European ancestry living in Framingham, Massachusetts. The COPDGene study recruited non-Hispanic African American and White participants aged 45 to 80 years with a smoking history of 10 pack-years or more from 21 US clinical centers. Participants who reported no previous physician diagnosis of COPD, emphysema, or chronic bronchitis were classified as having no history of COPD. The sample selection of each study is shown in Figure 1. We analyzed data from the FHS Offspring Cohort Exam 9, collected between 2011 and 2014, and Generation 3 Cohort Exam 2, collected between 2008 and 2011. Data from the COPDGene study baseline visit were collected between 2007 and 2012. We performed analyses from October 2023 to August 2024. The FHS and COPDGene studies were approved by the institutional review boards at each participating institution, and all study participants provided written informed consent.

Figure 1. Study Flow of the Analytic Samples From the Framingham Heart Study and COPDGene Study.

Figure 1.

Open in a new tab

aThe mLFQ score was calculated based on study participants’ data that best matched the questions from the Lung Function Questionnaire, a COPD case-finding questionnaire using age, smoking history, and respiratory symptoms to estimate an individual’s clinical risk for COPD. Further details of the calculation of the mLFQ score can be found in the eMethods in Supplement 1.

bGOLD 1 COPD is defined as forced expiratory volume in the first second/forced vital capacity (FEV1/FVC) <0.7 and FEV1 (percent predicted) ≥80%.21

cPRISm is defined as FEV1/FVC ≥0.7 and FEV1 (percent predicted) <80%.22

COPD indicates chronic obstructive pulmonary disease; GOLD, Global Initiative for Chronic Obstructive Lung Disease; mLFQ, modified Lung Function Questionnaire; PRISm, preserved ratio impaired spirometry.

COPD PRS

The COPD PRS was developed based on the UK Biobank and SpiroMeta (participants predominantly of European ancestry) genome-wide association study meta-analysis summary statistics for forced expiratory volume in the first second of expiration (FEV1) and FEV1/forced vital capacity (FVC), which has been previously described.16 A COPD PRS was calculated as a weighted sum of the PRSs for FEV1 and FEV1/FVC for each participant and standardized to a mean of 0 and an SD of 1 within each study cohort. Neither FHS nor COPDGene data were used to develop the PRS.

Modified Lung Function Questionnaire Score

We calculated a COPD clinical risk score using the Lung Function Questionnaire (LFQ). Developed and validated externally,23,24,25 the LFQ assesses 5 domains: age category, smoking history category, and the qualitative severity of productive cough, wheezing, and dyspnea on exertion. Each answer is assigned a score from 1 to 5, with a lower score indicating older age, more cumulative smoking, or worse symptoms. Thus, the total LFQ score ranges from 5 to 25, with a lower score suggesting a higher clinical risk for COPD. We calculated a modified LFQ (mLFQ) score for study participants based on their age, smoking history, and self-reported respiratory symptoms by selecting and coding responses from questions that best matched the LFQ questions.

Statistical Analysis

We conducted cross-sectional analyses using the PRS and mLFQ score as exposure variables, with spirometry-defined COPD as the outcome, all obtained during the same study visit. We constructed 3 logistic regression models with predictive variables of PRS (model 1), mLFQ score (model 2), and PRS plus mLFQ score (model 3) for the outcome of moderate to severe COPD, which was defined as FEV1/FVC less than 0.7 and FEV1 (percent predicted) less than 80%.21 Controls were defined as individuals with FEV1/FVC greater than or equal to 0.7 and FEV1 (percent predicted) greater than or equal to 80%. All analyses used principal components to adjust for genetic ancestry and used generalized estimating equations to account for known familial relatedness in FHS.

We evaluated model discrimination using the receiver operating characteristic curve and the area under the curve (AUC) metrics. We compared the AUCs of model 2 and model 3 using the DeLong test.26 We calculated the specificity, sensitivity, positive predictive value, and negative predictive value of model 3 using predefined exploratory risk thresholds of 5% (ie, spirometry referrals for individuals with a predicted COPD risk of 5% or higher), 10%, and 20%. We examined the outcome reclassification of model 3 compared with model 2 by calculating reclassification tables and categorical net reclassification indices for the risk thresholds.

We assessed model calibration using calibration plots and the Hosmer-Lemeshow test.27 We also calculated the integrated calibration index for the prediction models, with a lower integrated calibration index indicating better calibration, ie, a smaller mean difference between the predicted and observed COPD risk. Using decision curves,28 we examined the net benefit of prediction model–guided spirometry referral and the universal screening strategies compared with no screening spirometry, where net benefit measures additional COPD cases identified, adjusting for the relative harm/cost of negative spirometry tests.

Subgroup Analysis

To assess whether the PRS performs better in specific subgroups, we evaluated effect modification by age, sex, and pack-years of smoking and conducted stratified analyses by age group (younger [<55 y] vs older [≥55 y]), sex, and ever-smoking status. Further details of the study cohorts, the PRS, the mLFQ score, and statistical methods are provided in the eMethods in Supplement 1.

Results

Baseline Characteristics

A total of 3385 FHS and 4095 COPDGene participants who did not report a history of physician-diagnosed COPD were included (Table 1). Among these, 160 (4.7%) FHS and 775 (18.9%; 297 non-Hispanic African American [hereafter referred to as African American] and 478 non-Hispanic White [hereafter referred to as “White”]) COPDGene participants were found to have spirometry-defined moderate to severe COPD. Compared with the community-based FHS participants, COPDGene participants were more often males and had a greater smoking history and a higher prevalence of respiratory symptoms. Compared with controls, participants with COPD were older and had a greater cumulative smoking history and increased respiratory symptoms, which was reflected in lower (worse) mLFQ scores (eTable 1 in Supplement 1). Participants with COPD also had a higher prevalence of self-reported asthma history and higher COPD PRS.

Table 1. Baseline Characteristics of the Framingham Heart Study and the Genetic Epidemiology of COPD (COPDGene) Study Participantsa.

Characteristics Framingham Heart Study (n = 3385) COPDGene study (n = 4095)
Demographics
Age, median (IQR), y 52.0 (45.0-63.9) 56.8 (50.7-63.9)
Sex, No. (%)
Male 1553 (45.9) 2273 (55.5)
Female 1832 (54.1) 1822 (44.5)
Race and ethnicity, No. (%)
African American 0 1343 (32.8)
Hispanic or Latino 2 (0.1)
White 3336 (98.6) 2752 (67.2)
Otherb 45 (1.3)
Smoking history
Smoking status, No. (%)
Never 1801 (53.2)
Former 1365 (40.3) 1830 (44.7)
Current 219 (6.5) 2265 (55.3)
Pack-years among those who smoked, median (IQR) 10.0 (3.3-21.0) 35.0 (23.4-47.2)
Respiratory symptom, No. (%)c
Dyspnea 236 (7.0) 1417 (34.6)
Cough 218 (6.4) 1115 (27.2)
Phlegm 209 (6.2) 1116 (27.3)
Wheezing 530 (15.7) 1226 (29.9)
Self-reported history of asthma, No. (%) 190 (5.6) 207 (5.1)
mLFQ score, median (IQR)d 22.0 (21.0-23.0) 18.0 (16.0-19.0)
Polygenic risk score, mean (SD)e 0 (1.0) 0 (1.0)
Lung function
FEV1 (percent predicted), median (IQR), % 106.8 (97.8-116.3) 93.2 (83.0-104.4)
FEV1/FVC, median (IQR) 0.77 (0.74-0.80) 0.77 (0.72-0.82)
Moderate to severe COPD, No. (%)f 160 (4.7) 775 (18.9)
Open in a new tab

Abbreviations: COPD, chronic obstructive pulmonary disease; FEV1, forced expiratory volume in the first second of expiration; FVC, forced vital capacity; mLFQ, modified Lung Function Questionnaire.

a

The Framingham Heart Study enrolled community-dwelling individuals predominantly of European ancestry living in Framingham, Massachusetts. The COPDGene Study enrolled non-Hispanic African American and White participants with a smoking history of 10 pack-years or more from 21 centers in the US.

b

Self-reported race and ethnicity questionnaires were administered during several Framingham Heart Study examinations, and participants may have responded more than once. In later examinations, participants were allowed to select more than 1 race. “Other” indicates that a participant either preferred not to answer or selected multiple races.

c

The presence (vs absence) of a respiratory symptom.

d

The mLFQ score was calculated using study participants’ data that best matched the questions from the Lung Function Questionnaire, a COPD case-finding questionnaire using age, smoking history, and respiratory symptoms to estimate an individual’s clinical risk for COPD. The mLFQ score is designed to range from 5 to 25, with a lower score indicating a higher clinical risk for COPD.

e

The COPD polygenic risk score was standardized to a mean of 0 and an SD of 1 within each analytic sample of the Framingham Heart Study, the COPDGene African American, and COPDGene White participants.

f

Moderate to severe COPD is defined as FEV1/FVC less than 0.7 and FEV1 (percent predicted) less than 80%.

Associations of PRS and mLFQ Score With COPD

In the multivariable logistic regression analyses, a 1-unit increase in the SD of the PRS was associated with moderate to severe COPD in FHS (odds ratio [OR], 2.34 [95% CI, 1.94-2.82]), COPDGene African American (OR, 1.49 [95% CI, 1.30-1.72]), and COPDGene White (OR, 1.79 [95% CI, 1.60-2.01]) participants without a history of COPD. Additionally, every 1-unit decrease in the mLFQ score was associated with COPD in the FHS (OR, 1.48 [95% CI, 1.40-1.58]), COPDGene African American (OR, 1.19 [95% CI, 1.14-1.23]), and COPDGene White (OR, 1.34 [95% CI, 1.30-1.39]) participants.

Model Performance

For discrimination performance, the receiver operating characteristic and AUC of prediction models are shown in Figure 2. Adding the PRS to the mLFQ score significantly improved the AUC from 0.78 to 0.84 (change in AUC, 0.06 [95% CI, 0.03-0.09]; P < .001) for FHS, from 0.69 to 0.72 (change in AUC, 0.02 [95% CI, 0.001 to 0.04]; P = .04) for COPDGene African American, and from 0.75 to 0.78 (change in AUC, 0.03 [95% CI, 0.01 to 0.05]; P < .001) for COPDGene White participants. The sensitivity, specificity, positive predictive value, and negative predictive value of the combined PRS and mLFQ score for predicting COPD using risk thresholds of 5%, 10%, and 20% are shown in Table 2.

Figure 2. Receiver Operating Characteristic Curve and Area Under the Curve for the PRS, mLFQ Score, and Combined PRS and mLFQ Score in Predicting Moderate to Severe COPD.

Figure 2.

Open in a new tab

Moderate to severe COPD is defined as forced expiratory volume in the first second/forced vital capacity (FEV1/FVC) <0.7 and FEV1 (percent predicted) <80%. The mLFQ score was calculated using study participants’ data that best matched the questions from the Lung Function Questionnaire, a COPD case-finding questionnaire using age, smoking history, and respiratory symptoms to estimate an individual’s clinical risk for COPD. AUC indicates area under the receiver operating characteristic curve; COPD, chronic obstructive pulmonary disease; COPDGene, Genetic Epidemiology of Chronic Obstructive Pulmonary Disease study; mLFQ, modified Lung Function questionnaire; PRS, polygenic risk score.

Table 2. Performance Measures of the Combined PRS and mLFQ Score in Predicting Moderate to Severe COPD at Predefined Risk Thresholds.

Risk threshold, %a Result, No. of participants Performance metric (95% CI)
True-positive False-negative False-positive True-negative Sensitivity Specificity PPVb NPVb
Framingham Heart Study participants
≥5 118 42 699 2526 0.74 (0.66-0.80) 0.79 (0.78-0.80) 0.15 (0.13-0.18) 0.98 (0.98-0.99)
≥10 87 73 282 2943 0.54 (0.46-0.62) 0.91 (0.90-0.92) 0.24 (0.19-0.28) 0.98 (0.97-0.98)
≥20 40 120 89 3136 0.25 (0.19-0.32) 0.97 (0.97-0.98) 0.31 (0.23-0.40) 0.96 (0.96-0.97)
COPDGene White participants
≥5 470 8 1864 410 0.98 (0.97-0.99) 0.18 (0.16-0.20) 0.20 (0.19-0.22) 0.98 (0.96-0.99)
≥10 411 67 1158 1116 0.86 (0.83-0.89) 0.49 (0.47-0.51) 0.26 (0.24-0.28) 0.94 (0.93-0.96)
≥20 288 190 475 1799 0.60 (0.56-0.65) 0.79 (0.77-0.81) 0.38 (0.34-0.41) 0.90 (0.89-0.92)
COPDGene African American participants
≥5 297 0 1031 15 1.00 (0.99-1.00) 0.01 (0.01-0.02) 0.22 (0.20-0.25) 1.00 (0.78-1.00)
≥10 280 17 856 190 0.94 (0.91-0.97) 0.18 (0.16-0.21) 0.25 (0.22-0.27) 0.92 (0.87-0.95)
≥20 210 87 387 659 0.71 (0.65-0.76) 0.63 (0.60-0.66) 0.35 (0.31-0.39) 0.88 (0.86-0.91)
Open in a new tab

Abbreviations: COPD, chronic obstructive pulmonary disease; FEV1/FVC, forced expiratory volume in the first second of expiration/forced vital capacity; mLFQ, modified Lung Function Questionnaire; NPV, negative predictive value; PPV, positive predictive value; PRS, polygenic risk score.

a

A participant is predicted to have COPD if the combined PRS and mLFQ score predicted risk is greater than or equal to the predefined risk threshold. The predicted COPD status is compared with the observed spirometry-defined moderate to severe COPD. Moderate-to-severe COPD is defined as FEV1/FVC <0.7 and FEV1 (percent predicted) <80%. The mLFQ score was calculated using study participants’ data that best matched the questions from the Lung Function Questionnaire, a COPD case-finding questionnaire using age, smoking history, and respiratory symptoms to estimate an individual’s clinical risk for COPD.

b

PPV and NPV are calculated based on the observed prevalence of undiagnosed COPD: 4.7% in the Framingham Heart Study and 18.9% in COPDGene (17.4% in COPDGene White participants and 22.1% in COPDGene African American participants).

For reclassification performance, using a risk threshold of 10%, the addition of the PRS to the mLFQ score correctly reclassified 13.8% (95% CI, 6.6%-21.0%) of participants (22 of 160) with COPD in FHS; in contrast, adding the PRS to the mLFQ score did not reclassify cases, but did correctly reclassify controls as not having COPD in COPDGene (Figure 3). Using alternative risk thresholds of 5% (eFigure 1 in Supplement 1) and 20% (eFigure 2 in Supplement 1), similar net reclassifications of cases were observed in FHS. In COPDGene, adding the PRS to the mLFQ score correctly reclassified COPD cases using a risk threshold of 20%.

Figure 3. Net Reclassification Comparing the Combined PRS and mLFQ Score vs the mLFQ Score Alone for Predicting Moderate to Severe COPD Using a Risk Threshold ≥10%.

Figure 3.

Open in a new tab

Moderate to severe COPD is defined as forced expiratory volume in the first second/forced vital capacity (FEV1/FVC) <0.7 and FEV1 (percent predicted) <80%. The modified Lung Function questionnaire (mLFQ) score was calculated using study participants’ data that best matched the questions from the Lung Function Questionnaire, a chronic obstructive pulmonary disease (COPD) case finding questionnaire using age, smoking history, and respiratory symptoms to estimate an individual’s clinical risk for COPD. A model classifies an individual as a case if the predicted risk is higher than or equal to the risk threshold. The net reclassification index (NRI) quantifies how well a model (in this case, the combined polygenic risk score [PRS] and mLFQ score) reclassifies individuals into correct risk categories compared with an older model (mLFQ score alone). The NRI for cases is calculated as the proportion of cases correctly reclassified into a higher risk category by the new model minus the proportion of cases incorrectly reclassified into a lower risk category. Similarly, the NRI for noncases is calculated as the percentage of noncases correctly reclassified into a lower risk category by the new model minus the percentage of noncases incorrectly reclassified into a higher risk category. The net reclassification comparing the combined PRS and mLFQ score vs the mLFQ score alone, using risk thresholds of ≥5% and ≥20%, is provided in eFigures 1 and 2 in Supplement 1, respectively. COPDGene indicates Genetic Epidemiology of COPD Study.

The calibration plots of the prediction models are shown in eFigure 3 in Supplement 1. A significant difference between the mLFQ score predicted risk and the observed risk (ie, miscalibration) was found in COPDGene White participants (Hosmer-Lemeshow test P value = .03), but not in FHS or COPDGene African American participants. There was no evidence of miscalibration of the PRS or the combined PRS and mLFQ score predicted risk among study cohorts. The integrated calibration indices of the prediction models are shown in eTable 2 in Supplement 1.

In decision curve analyses, the combined PRS and mLFQ score showed a positive and higher net benefit compared with the mLFQ score alone and universal screening spirometry across a range of risk thresholds in FHS and COPDGene; however, the relative and absolute magnitude of the incremental net benefit of the PRS varied depending on the risk threshold and the study cohort (eFigure 4 in Supplement 1).

Subgroup Analyses

In the logistic regression models for predicting spirometry-defined COPD, a significant interaction was detected between the PRS and age (β for interaction = −0.014; P = .04) in FHS, but not for sex and pack-years. In stratified analyses, the PRS demonstrated better discrimination performance in the younger group (<55 y; AUC, 0.80 [95% CI, 0.74-0.87]) compared with the older group (age ≥55; AUC, 0.70 [95% CI, 0.65-0.75]) in FHS. Additional results are provided in eResults and eTable 3 in Supplement 1.

Discussion

In this observational study using US-based cohorts of community-dwelling adults and COPD-enriched people who smoked without previously diagnosed COPD, it was demonstrated that a COPD PRS enhances the prediction of moderate to severe COPD beyond a case-finding questionnaire. The findings suggest that an individual’s genetic risk for COPD, as indicated by a PRS, has a potential added value to conventional case-finding approaches for identifying undiagnosed COPD and guiding referrals for confirmatory spirometry, particularly in community-based settings.

Approximately 5% of adults in the community-based FHS had undiagnosed moderate-to-severe COPD, representing many individuals with undiagnosed COPD in the community. The US Preventive Services Task Force recommends against universal screening spirometry in asymptomatic adults partly due to concerns regarding cost-effectiveness, which may be mitigated by case-finding approaches. In FHS, a positive net benefit of a questionnaire-based case-finding method (the modified Lung Function Questionnaire [mLFQ] score) was found across various risk thresholds of spirometry testing. It was also shown that adding the COPD PRS to the mLFQ score identified additional COPD cases across a range of risk thresholds in FHS participants. The combined PRS and mLFQ score demonstrated higher sensitivity at lower risk thresholds and higher positive predictive value at higher risk thresholds, highlighting the importance of considering the appropriate threshold for spirometry referrals in specific case-finding environments.

In contrast, among COPDGene participants who had more severe smoking history, adding the PRS to the mLFQ score identified additional COPD cases only at relatively high spirometry testing thresholds, and universal screening spirometry showed comparable net benefit to the case-finding approaches at relatively low risk thresholds. These findings suggest that the benefits of the PRS depend on the prevalence of undiagnosed COPD and the setting in which a spirometry testing threshold is deemed appropriate.

COPD is a heterogeneous condition influenced by gene-environment interactions throughout the lifespan.29 The COPD PRS is associated with lung function growth patterns and may at least partially represent developmental origins of COPD.16 Thus, the PRS may identify COPD cases of distinct pathogenetic pathways compared with cases identified based on nongenetic risk factors. In FHS, it was observed that the PRS demonstrated superior performance in identifying undiagnosed COPD in younger compared with older adults, highlighting the potential role of genetic risk in the early diagnosis of COPD and consistent with previous work demonstrating that the PRS was associated with the development of COPD in young adults, suggesting that the PRS could guide preventive measures for those at high genetic risk.17

Although the calculation of PRS requires genetic testing, this testing can be done before the onset of symptoms or lung function impairment and only needs to be done once in a person’s lifetime. The cost of whole-genome genotyping can be $35 or less per person, and these data can be used to inform genetic risks of multiple diseases.30 A potentially practical approach would be to apply the PRS in the case finding of COPD among individuals who already have their genome genotyped for other reasons.

COPD was defined using spirometry criteria because airflow obstruction is the hallmark of COPD and to maintain consistency with previous case-finding studies. However, the diagnosis of COPD may involve features beyond airflow obstruction.31 Moderate to severe COPD was used as the primary outcome because it identifies a group with currently available evidence-based therapies and is less likely to be misclassified compared with milder disease. Mild COPD and preserved ratio impaired spirometry were not included due to the current lack of effective evidence-based treatments. Additionally, while preserved ratio impaired spirometry is sometimes considered a pre-COPD condition, it is a nonspecific spirometry phenotype that represents heterogeneous underlying pathophysiology beyond airflow obstruction.32

The COPD PRS captures only a small proportion of COPD heritability. Although significant results were observed in the COPDGene African American cohort, the predictive ability of the PRS was lower in this cohort compared with White participants. This reflects the well-known issue of low trans-ancestry portability of PRS,33,34 as the COPD PRS was developed based on genome-wide association studies using samples from individuals of European ancestry. The development of multiancestry PRS, along with advancements in PRS construction methods, may improve the predictive performance and the portability of the PRS in the future.35

Among the COPD case-finding questionnaires, the validated LFQ was chosen to calculate a mLFQ score for study participants, primarily because the LFQ was developed using a representative US sample and overlaps with the questionnaires used in previous studies. Additionally, the LFQ performs comparably to other commonly used questionnaires.12,36 Handheld airflow measurement devices, such as peak flow meters and microspirometers, have been shown to facilitate COPD diagnosis in primary care settings, especially when combined with a questionnaire (AUC, 0.76-0.95).12 Despite the reported high performance of handheld devices, their use often requires a medical office visit. Additionally, microspirometers may have technical requirements and costs that make them more suitable for high-income countries.37 Handheld device–based airflow measurements were not available in the current study cohorts and therefore it was not possible to compare the performance of the PRS with handheld devices or evaluate the added utility of the PRS in conjunction with these case-finding measures. The combined PRS and questionnaire may offer complementary value to handheld devices and warrant further evaluation in future studies.

Limitations

This study has limitations. First, risk thresholds used to assess model performance, although reasonable, were not evidence-based because there is no well-accepted risk cutoff in practice. Second, a new risk prediction equation or weighted scoring tool combining the PRS with the conventional case-finding approach were not developed. The primary goal of this study was to address the knowledge gap regarding whether a COPD PRS adds value to contemporary case-finding tools. Third, prebronchodilator spirometry was used to define COPD in the FHS, which may have ascertained individuals with reversible moderate to severe airflow obstruction. However, a significant percentage of patients with COPD exhibit bronchodilator responsiveness, which is associated with a higher symptom burden and lower lung function38; thus, identifying these individuals could also be clinically relevant. Fourth, individuals with very severe COPD were less likely to participate in the FHS; however, they may have already been diagnosed by a physician. Fifth, the generalizability of the US cohort-based results to people in other countries, especially low- and middle-income countries, needs further studies. Sixth, it is not known whether the additional COPD cases identified by a PRS will lead to improved clinical outcomes or whether such an approach is cost-effective.

Conclusions

A COPD polygenic risk score enhances the identification of adults with undiagnosed COPD beyond a case-finding questionnaire based on clinical risk factors and respiratory symptoms, especially in community-based settings. Although the findings suggest a potential role for genetic risk in COPD case finding, further research is needed to evaluate its clinical use in improving diagnosis and outcomes before it can be applied in practice.

Supplement 1.

eMethods

eResults

eTable 1. Baseline Characteristics of Study Participants, Stratified by the Presence of Spirometry-Defined Moderate-to-Severe COPD

eTable 2. Integrated Calibration Index (ICI) of the Prediction Models

eTable 3. AUC of the PRS, the mLFQ Score, and the Combined PRS and mLFQ Score for Predicting Moderate-to-Severe COPD, Stratified by Age Group, Sex, and Ever Smoking Status

eFigure 1. Net Reclassification Comparing the Combined PRS and mLFQ Score Versus the mLFQ Score Alone for Predicting Moderate-to-Severe COPD Using a Risk Threshold of ≥ 5%

eFigure 2. Net Reclassification Comparing the Combined PRS and mLFQ Score Versus the mLFQ Score Alone for Predicting Moderate-to-Severe COPD Using a Risk Threshold of ≥ 20%

eFigure 3. Calibration Plots of the PRS, the mLFQ Score, and the Combined PRS and mLFQ Score for Predicting Moderate-to-Severe COPD

eFigure 4. Net Benefit of Spirometry Referrals Guided by the PRS, the mLFQ Score, the Combined PRS and mLFQ Score, and the Universal Screening Approach Compared to No Screening Spirometry

eReferences

jama-e2424212-s001.pdf (1.2MB, pdf)
Supplement 2.

Data sharing statement

jama-e2424212-s002.pdf (15.7KB, pdf)

References

  • 1.Boers E, Barrett M, Su JG, et al. Global burden of chronic obstructive pulmonary disease through 2050. JAMA Netw Open. 2023;6(12):e2346598. doi: 10.1001/jamanetworkopen.2023.46598 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Lamprecht B, Soriano JB, Studnicka M, et al. ; BOLD Collaborative Research Group, the EPI-SCAN Team, the PLATINO Team, and the PREPOCOL Study Group . Determinants of underdiagnosis of COPD in national and international surveys. Chest. 2015;148(4):971-985. doi: 10.1378/chest.14-2535 [DOI] [PubMed] [Google Scholar]
  • 3.Martinez CH, Mannino DM, Jaimes FA, et al. Undiagnosed obstructive lung disease in the United States: associated factors and long-term mortality. Ann Am Thorac Soc. 2015;12(12):1788-1795. doi: 10.1513/AnnalsATS.201506-388OC [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Çolak Y, Afzal S, Nordestgaard BG, Vestbo J, Lange P. Prognosis of asymptomatic and symptomatic, undiagnosed COPD in the general population in Denmark: a prospective cohort study. Lancet Respir Med. 2017;5(5):426-434. doi: 10.1016/S2213-2600(17)30119-4 [DOI] [PubMed] [Google Scholar]
  • 5.Labonté LE, Tan WC, Li PZ, et al. ; Canadian Respiratory Research Network; CanCOLD Collaborative Research Group . Undiagnosed chronic obstructive pulmonary disease contributes to the burden of health care use: data from the CanCOLD study. Am J Respir Crit Care Med. 2016;194(3):285-298. doi: 10.1164/rccm.201509-1795OC [DOI] [PubMed] [Google Scholar]
  • 6.Gerstein E, Bierbrier J, Whitmore GA, et al. Impact of undiagnosed chronic obstructive pulmonary disease and asthma on symptoms, quality of life, healthcare use, and work productivity. Am J Respir Crit Care Med. 2023;208(12):1271-1282. doi: 10.1164/rccm.202307-1264OC [DOI] [PubMed] [Google Scholar]
  • 7.Jordan RE, Lam KB, Cheng KK, et al. Case finding for chronic obstructive pulmonary disease: a model for optimising a targeted approach. Thorax. 2010;65(6):492-498. doi: 10.1136/thx.2009.129395 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Johnson KM, Sadatsafavi M, Adibi A, et al. Cost effectiveness of case detection strategies for the early detection of COPD. Appl Health Econ Health Policy. 2021;19(2):203-215. doi: 10.1007/s40258-020-00616-2 [DOI] [PubMed] [Google Scholar]
  • 9.Jordan RE, Adab P, Sitch A, et al. Targeted case finding for chronic obstructive pulmonary disease versus routine practice in primary care (TargetCOPD): a cluster-randomised controlled trial. Lancet Respir Med. 2016;4(9):720-730. doi: 10.1016/S2213-2600(16)30149-7 [DOI] [PubMed] [Google Scholar]
  • 10.Lambe T, Adab P, Jordan RE, et al. Model-based evaluation of the long-term cost-effectiveness of systematic case-finding for COPD in primary care. Thorax. 2019;74(8):730-739. doi: 10.1136/thoraxjnl-2018-212148 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Martinez FJ, Mannino D, Leidy NK, et al. ; High-Risk-COPD Screening Study Group * . A new approach for identifying patients with undiagnosed chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 2017;195(6):748-756. doi: 10.1164/rccm.201603-0622OC [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Aaron SD, Montes de Oca M, Celli B, et al. Early diagnosis and treatment of chronic obstructive pulmonary disease: the costs and benefits of case finding. Am J Respir Crit Care Med. 2024;209(8):928-937. doi: 10.1164/rccm.202311-2120PP [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Aaron SD, Vandemheen KL, Whitmore GA, et al. ; UCAP Investigators . Early diagnosis and treatment of COPD and asthma: a randomized, controlled trial. N Engl J Med. 2024;390(22):2061-2073. doi: 10.1056/NEJMoa2401389 [DOI] [PubMed] [Google Scholar]
  • 14.Zhou JJ, Cho MH, Castaldi PJ, Hersh CP, Silverman EK, Laird NM. Heritability of chronic obstructive pulmonary disease and related phenotypes in smokers. Am J Respir Crit Care Med. 2013;188(8):941-947. doi: 10.1164/rccm.201302-0263OC [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Kullo IJ, Lewis CM, Inouye M, Martin AR, Ripatti S, Chatterjee N. Polygenic scores in biomedical research. Nat Rev Genet. 2022;23(9):524-532. doi: 10.1038/s41576-022-00470-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Moll M, Sakornsakolpat P, Shrine N, et al. ; International COPD Genetics Consortium; SpiroMeta Consortium . Chronic obstructive pulmonary disease and related phenotypes: polygenic risk scores in population-based and case-control cohorts. Lancet Respir Med. 2020;8(7):696-708. doi: 10.1016/S2213-2600(20)30101-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Zhang J, Xu H, Qiao D, et al. A polygenic risk score and age of diagnosis of COPD. Eur Respir J. 2022;60(3):2101954. doi: 10.1183/13993003.01954-2021 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Kannel WB, Feinleib M, McNamara PM, Garrison RJ, Castelli WP. An investigation of coronary heart disease in families: the Framingham Offspring Study. Am J Epidemiol. 1979;110(3):281-290. doi: 10.1093/oxfordjournals.aje.a112813 [DOI] [PubMed] [Google Scholar]
  • 19.Splansky GL, Corey D, Yang Q, et al. The third generation cohort of the National Heart, Lung, and Blood Institute’s Framingham Heart Study: design, recruitment, and initial examination. Am J Epidemiol. 2007;165(11):1328-1335. doi: 10.1093/aje/kwm021 [DOI] [PubMed] [Google Scholar]
  • 20.Regan EA, Hokanson JE, Murphy JR, et al. Genetic epidemiology of COPD (COPDGene) study design. COPD. 2010;7(1):32-43. doi: 10.3109/15412550903499522 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Global Initiative for Chronic Obstructive Lung Disease (GOLD). Global Strategy for Prevention, Diagnosis, and Management of COPD: 2024. Report. Accessed September 29, 2024. https://goldcopd.org/2024-gold-report/
  • 22.Wan ES, Castaldi PJ, Cho MH, et al. ; COPDGene Investigators . Epidemiology, genetics, and subtyping of preserved ratio impaired spirometry (PRISm) in COPDGene. Respir Res. 2014;15(1):89. doi: 10.1186/s12931-014-0089-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Yawn BP, Mapel DW, Mannino DM, et al. ; Lung Function Questionnaire Working Group . Development of the Lung Function Questionnaire (LFQ) to identify airflow obstruction. Int J Chron Obstruct Pulmon Dis. 2010;5:1-10. [PMC free article] [PubMed] [Google Scholar]
  • 24.Hanania NA, Mannino DM, Yawn BP, et al. Predicting risk of airflow obstruction in primary care: Validation of the lung function questionnaire (LFQ). Respir Med. 2010;104(8):1160-1170. doi: 10.1016/j.rmed.2010.02.009 [DOI] [PubMed] [Google Scholar]
  • 25.Spyratos D, Haidich AB, Chloros D, Michalopoulou D, Sichletidis L. Comparison of three screening questionnaires for chronic obstructive pulmonary disease in the primary care. Respiration. 2017;93(2):83-89. doi: 10.1159/000453586 [DOI] [PubMed] [Google Scholar]
  • 26.DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics. 1988;44(3):837-845. doi: 10.2307/2531595 [DOI] [PubMed] [Google Scholar]
  • 27.Hosmer DW, Lemesho S. Further work on a goodness of fit test for the multiple logistic regression-model. Biometrics. 1979;35(4):889-889. doi: 10.1080/03610928008827941 [DOI] [Google Scholar]
  • 28.Vickers AJ, Elkin EB. Decision curve analysis: a novel method for evaluating prediction models. Med Decis Making. 2006;26(6):565-574. doi: 10.1177/0272989X06295361 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Agustí A, Melén E, DeMeo DL, Breyer-Kohansal R, Faner R. Pathogenesis of chronic obstructive pulmonary disease: understanding the contributions of gene-environment interactions across the lifespan. Lancet Respir Med. 2022;10(5):512-524. doi: 10.1016/S2213-2600(21)00555-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.PGS Catalog. The Polygenic Score (PGS) Catalog . Accessed August 4, 2024. https://www.pgscatalog.org/
  • 31.Lowe KE, Regan EA, Anzueto A, et al. COPDGene 2019: redefining the diagnosis of chronic obstructive pulmonary disease. Chronic Obstr Pulm Dis. 2019;6(5):384-399. doi: 10.15326/jcopdf.6.5.2019.0149 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Wan ES, Hokanson JE, Regan EA, et al. Significant spirometric transitions and preserved ratio impaired spirometry among ever smokers. Chest. 2022;161(3):651-661. doi: 10.1016/j.chest.2021.09.021 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Duncan L, Shen H, Gelaye B, et al. Analysis of polygenic risk score usage and performance in diverse human populations. Nat Commun. 2019;10(1):3328. doi: 10.1038/s41467-019-11112-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Privé F, Aschard H, Carmi S, et al. Portability of 245 polygenic scores when derived from the UK Biobank and applied to 9 ancestry groups from the same cohort. Am J Hum Genet. 2022;109(1):12-23. doi: 10.1016/j.ajhg.2021.11.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Wang Y, Tsuo K, Kanai M, Neale BM, Martin AR. Challenges and opportunities for developing more generalizable polygenic risk scores. Annu Rev Biomed Data Sci. 2022;5:293-320. doi: 10.1146/annurev-biodatasci-111721-074830 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Schnieders E, Ünal E, Winkler V, et al. Performance of alternative COPD case-finding tools: a systematic review and meta-analysis. Eur Respir Rev. 2021;30(160):200350. doi: 10.1183/16000617.0350-2020 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Lin CH, Cheng SL, Chen CZ, Chen CH, Lin SH, Wang HC. Current progress of COPD early detection: key points and novel strategies. Int J Chron Obstruct Pulmon Dis. 2023;18:1511-1524. doi: 10.2147/COPD.S413969 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Beasley R, Hughes R, Agusti A, et al. Prevalence, diagnostic utility and associated characteristics of bronchodilator responsiveness. Am J Respir Crit Care Med. 2024;209(4):390-401. doi: 10.1164/rccm.202308-1436OC [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplement 1.

eMethods

eResults

eTable 1. Baseline Characteristics of Study Participants, Stratified by the Presence of Spirometry-Defined Moderate-to-Severe COPD

eTable 2. Integrated Calibration Index (ICI) of the Prediction Models

eTable 3. AUC of the PRS, the mLFQ Score, and the Combined PRS and mLFQ Score for Predicting Moderate-to-Severe COPD, Stratified by Age Group, Sex, and Ever Smoking Status

eFigure 1. Net Reclassification Comparing the Combined PRS and mLFQ Score Versus the mLFQ Score Alone for Predicting Moderate-to-Severe COPD Using a Risk Threshold of ≥ 5%

eFigure 2. Net Reclassification Comparing the Combined PRS and mLFQ Score Versus the mLFQ Score Alone for Predicting Moderate-to-Severe COPD Using a Risk Threshold of ≥ 20%

eFigure 3. Calibration Plots of the PRS, the mLFQ Score, and the Combined PRS and mLFQ Score for Predicting Moderate-to-Severe COPD

eFigure 4. Net Benefit of Spirometry Referrals Guided by the PRS, the mLFQ Score, the Combined PRS and mLFQ Score, and the Universal Screening Approach Compared to No Screening Spirometry

eReferences

jama-e2424212-s001.pdf (1.2MB, pdf)
Supplement 2.

Data sharing statement

jama-e2424212-s002.pdf (15.7KB, pdf)

Articles from JAMA are provided here courtesy of American Medical Association

RESOURCES