Abstract
Background
There is substantial unexplained variability in the development of disease. A genetic risk score for COPD identifies individuals at markedly elevated risk of COPD; however, many high-genetic risk individuals do not develop disease. We sought to define genetic resilience in COPD by identifying and characterizing individuals who are resistant to their elevated genetic susceptibility.
Methods
We defined resilience to genetic risk (genetic resilience) as absence of airflow obstruction (FEV1/FVC ≥ 0.70) in individuals with cigarette smoking exposure with a polygenic risk score for COPD at the 90th percentile or above. We defined clinical resilience according to previously published criteria, including a low symptom burden, limited radiographic disease, and normal lung-function decline despite similar smoking history. Using data from the Genetic Epidemiology of COPD (COPDGene) study, we compared genetically resilient individuals to clinically resilient individuals and genetic risk-matched individuals with COPD on clinical characteristics, radiographic findings, longitudinal outcomes, mortality, biomarkers, and social determinants of health.
Results
We found that, after adjustment for covariates, genetically resilient individuals (n = 144) had better lung function (β = 35.9% predicted, p < 0.001), fewer symptoms, and less radiographic disease compared to genetic risk-matched individuals with COPD (n = 362). Conversely, when compared to clinically resilient individuals (n = 420), genetically resilient individuals had slightly lower lung function and slightly worse radiographic measures of disease. Both clinically and genetically resilient individuals had higher survival compared to genetic-risk matched cases (hazard ratios = 0.34 and 0.41 with p < 0.001 and p = 0.002, respectively). While the majority of genetically resilient individuals remained resilient across the 5-year and 10-year follow-up visits, a higher proportion of clinically resilient individuals remained resilient across the follow-up period. Non-Hispanic white genetically resilient individuals had higher social vulnerability across multiple measures compared to both clinically resilient individuals and genetic risk-matched individuals with COPD.
Discussion
Genetic resilience to COPD represents a unique subtype of smokers that is distinct from clinical resilience and has important disease-related differences from genetic risk-matched individuals with COPD. Future studies are needed to identify the underlying biological contributors to the multiple types of resilience in COPD.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12931-025-03380-w.
Introduction
Cigarette smoking is the primary environmental risk factor for COPD in the United States [1, 2]. However, only a minority of individuals that smoke develop COPD, with an estimated prevalence from 20% to 50% in elderly populations [3, 4]. In addition, there is marked heterogeneity in disease susceptibility and in clinical manifestations of COPD among individuals that smoke having a comparable burden of cigarette smoking [5]. Genetic factors are thought to account for some of the observed clinical heterogeneity in COPD [6].
The estimated heritability, or proportion of variation in risk accounted for by genetic differences, for COPD ranges from 37% to 50% [7]. Genome-wide association studies (GWAS) have found several single nucleotide polymorphisms (SNPs) that are associated with COPD, though the effect size of each individual SNP is small [8, 9]. A polygenic risk score (PRS), which summarizes the effects of SNPs across the genome, has been shown to improve prediction of COPD [10]. The development of the PRS represents an important advance in the understanding of the genetic risk of COPD and opens the door to studying genetic resilience in COPD as resilience can only be defined against a known background of heightened risk.
The concept of genetic resilience, which has been successfully demonstrated in both Mendelian diseases and complex disorders, focuses on studying individuals who remain unaffected despite having high genetic risk [11, 12]. The existence of individuals who smoke cigarettes and are resilient to developing COPD has been recognized for several decades [4]. Investigators from the SPIROMICS study developed a multi-dimensional clinical definition of resilience in COPD and found that such “clinically resilient” individuals accounted for a modest proportion of study participants [12]. However, the prevalence and clinical features of individuals that smoke who are genetically resilient to COPD is unknown.
In the Genetic Epidemiology of COPD (COPDGene) study, we used the previously published PRS for COPD to identify individuals that smoke unaffected with COPD but at high genetic risk (genetically resilient) and compared them to individuals that smoke with COPD [13]. We also compared the genetically resilient individuals to individuals that smoke who met the definition of clinical resilience to further refine the characteristics specific to genetic resilience in COPD. We sought to identify clinical, physiologic, and imaging characteristics that differ between genetically resilient individuals, genetic risk-matched individuals with COPD, and clinically resilient individuals. We hypothesized that individuals who are genetically resilient to COPD would have distinct clinical and biological features compared to risk-matched individuals with COPD and clinically resilient individuals.
Methods
Study participants
COPDGene is a multicenter longitudinal observational study (ClinicalTrials.gov Identifier: NCT00608764, Registration Date 06 Feb 2008). COPDGene was designed to identify genetic factors associated with COPD and includes 10,198 subjects enrolled at 21 centers in the United States. Further details regarding recruitment have been previously published [13]. We defined COPD according to the Global Initiative for Chronic Obstructive Lung Disease (GOLD) guidelines, as post-bronchodilator forced expiratory volume in one second (FEV1) to forced vital capacity (FVC) ratio < 0.70 [14]. Study participants were invited to participate in 5-year (Phase 2) and 10-year (Phase 3) follow-up visits. Study participants completed questionnaires to quantify symptom burden, a standard spirometry protocol, and chest computed tomography (CT) scans.
Polygenic risk score and definition of genetic and clinical resilience
Genetic risk for COPD in the COPDGene sample was quantified using the PRS developed by Moll and colleagues, which did not include the COPDGene data in its derivation. Briefly, SNP weights were generated using the estimated effect sizes from external GWASs for FEV1 and FEV1/FVC. The authors applied a penalized regression framework that accounted for linkage disequilibrium and created a combined polygenic risk score as a weighted sum of the scores for FEV1 and FEV1/FVC.
We defined individuals resilient to high genetic risk (genetically resilient) as those unaffected by COPD (i.e., FEV1/FVC ≥ 0.70) despite having a polygenic risk score above the 90th percentile (Fig. 1).
Fig. 1.
Polygenic risk score distribution among non-Hispanic white COPD case and control individuals in COPDGene. Dotted line denotes 90th percentile. COPD – chronic obstructive pulmonary disease; COPDGene – Genetic Epidemiology of COPD study
We defined risk-matched case individuals as those with COPD (i.e., FEV1/FVC < 0.70) and a PRS above the 90th percentile among all study participants [12]. We excluded individuals who met criteria for preserved ratio impaired spirometry (PRISm), i.e., FEV1/FVC ≥ 0.70 and FEV1 %predicted < 80%. We also performed sensitivity analyses using a secondary definition of genetic resilience, which included only those individuals who met the PRS threshold and had FEV1/FVC ≥ 0.70 at both the Phase 1 and Phase 2 visits, as well as a comparison of genetic resilience with individuals with COPD and low genetic risk (PRS below the 10th percentile).
We defined clinically resilient individuals as individuals with cigarette smoking exposure without COPD (i.e., FEV1/FVC ≥ 0.70) using the definition established in the SPIROMICS study, which includes: (1) cough and phlegm COPD Assessment Test (CAT) score < 2; (2) Modified Medical Research Council (mMRC) score 0–1; (3) no exacerbations in the 12 months prior to enrollment; (4) no exacerbations during the 3-year follow-up period; (5) percent emphysema less than the upper limit of normal generated from the Multi-Ethnic Study of Atherosclerosis by Hoffman et al. [15]; (6) air trapping (measured by PRMfSAD) below the upper limit of normal generated in SPIROMICS by Martinez et al. [16]; and (7) annual rate of decline in FEV1 less than the 95% upper limit confidence interval generated from the Framingham offspring cohort by Kohansal et al. [17]. We excluded individuals who met the criteria for both definitions of resilience.
Statistical analysis
We compared genetically resilient individuals to risk-matched individuals with COPD and clinically resilient individuals on demographics, comorbidities, lung function (including FEV1 percent predicted by the Global Lung Initiative [GLI] race-neutral equations), respiratory symptoms (as measured by the St. George’s Respiratory Questionnaire [SGRQ]), and chest CT scan measurements using Student’s t-tests for continuous variables and Chi-square tests for proportions. We repeated these comparisons after matching genetically resilient, genetic risk-matched, and clinically resilient individuals (1:1:1) on age, race, and smoking pack-years, using the MatchIt package.
We performed multivariable regression for a subset of outcome measures, including FEV1 percent predicted and CT scan measurements of emphysema (% LAA < −950 HU, Perc15) and airway pathology (Pi10), adjusted for age, gender, smoking pack-years, current smoking status, body mass index, the first five globally derived principal components of genetic ancestry, and CT scanner model for CT measurements. We performed additional sensitivity analyses after stratifying by race, further adjusted using ancestry-specific principal components [18].
We tested for differences in the proportion of individuals in each group that changed disease status at the 5-year follow-up visit (Phase 2) and the 10-year follow-up visit (Phase 3) using Chi-square tests for proportions. We tested for differences in change in FEV1 percent predicted and absolute value, St. George’s Respiratory Questionnaire (SGRQ) total score, six-minute walk distance and percent emphysema (only at Phase 2). We performed survival analysis using Kaplan-Meier curves and Cox proportional hazard models, adjusted for covariates as above as well as the BODE index, using the survival and survminer packages [19, 20].
To better understand the environmental contributors to each resilience subtype, we compared social determinants of health, measured by the Centers for Disease Control and Prevention and Agency for Toxic Substances and Disease Registry Social Vulnerability Index (SVI) components and overall percentile and the area deprivation index (ADI) [21]. SVI uses data across multiple domains from the U.S. Census Bureau to identify communities that are most likely to require support, from public health officials or emergency responders, in response to a hazardous event. ADI, on the other hand, ranks neighborhoods by relative socioeconomic conditions across the U.S. We compared the SVI components and overall percentile as well as ADI ranks using Student’s t-tests and performed multivariable regression analyses after stratifying by race and adjusting for educational attainment, income category, insurance status, and ancestry-specific principal component.
We also compared the expression of proteins in blood previously associated with COPD between genetically resilient individuals and risk-matched individuals with COPD and clinically resilient individuals [22–25]. The biomarkers were initially selected for correlation in the analysis to define clinical resilience. The details of proteomic measurement have been previously described [26, 27]. We compared inverse normalized levels of protein expression using Student’s t-test. All statistical analyses were performed using R (version 4.2.2), unless otherwise indicated.
Results
Baseline characteristics
The COPDGene Study includes 10,198 current and former individuals that smoke. We identified 144 individuals who met the definition of genetic resilience, 362 genetic risk-matched individuals with COPD, and 420 individuals who met the definition of clinical resilience (Table 1).
Table 1.
Subject characteristics by resilience subtype
| Characteristic | Genetic Resilience | Genetic Risk-Matched Cases | p | Clinical Resilience | p |
|---|---|---|---|---|---|
| n | 144 | 362 | 420 | ||
| Age, years | 57.1 (8.3) | 62.4 (8.2) | <0.001 | 58.8 (8.7) | 0.034 |
| Female sex (%) | 74 (51.4) | 185 (51.1) | 1.000 | 219 (52.1) | 0.952 |
| African-American race (%) | 47 (32.6) | 62 (17.1) | <0.001 | 77 (18.3) | 0.001 |
| Current smoker (%) | 67 (46.5) | 130 (35.9) | 0.035 | 155 (36.9) | 0.052 |
| Smoking pack-years | 34.9 (17.7) | 46.6 (22.3) | <0.001 | 35.2 (20.1) | 0.867 |
| PRS | 1.8 (0.4) | 1.8 (0.5) | 0.411 | -0.3 (0.9) | <0.001 |
| Body mass index, kg/m2 | 28.2 (5.0) | 28.7 (5.9) | 0.384 | 29.5 (5.6) | 0.013 |
| Six-minute walk distance, ft | 1571.3 (300.0) | 1310.1 (386.7) | <0.001 | 1625.4 (304.1) | 0.065 |
| Asthma (%) | 0.001 | 0.020 | |||
| No | 118 (81.9) | 233 (64.4) | 380 (90.5) | ||
| Yes | 19 (13.2) | 89 (24.6) | 27 (6.4) | ||
| Don’t know | 7 (4.9) | 40 (11.0) | 13 (3.1) | ||
| SGRQ total score | 14.5 (15.4) | 31.9 (22.0) | <0.001 | 8.5 (9.6) | <0.001 |
| BODE index | 0.5 (0.9) | 2.7 (2.3) | <0.001 | 0.1 (0.4) | <0.001 |
| FEV1 percent predicted | 94.6 (10.6) | 58.8 (21.7) | <0.001 | 97.1 (11.3) | 0.021 |
| Bronchodilator response (%) | 16 (11.1) | 125 (34.7) | <0.001 | 20 (4.8) | 0.013 |
| Percent emphysema | 2.3 (2.6) | 11.0 (11.8) | <0.001 | 2.0 (2.3) | 0.297 |
| Percent gas trapping | 11.6 (9.4) | 35.5 (19.4) | <0.001 | 8.9 (6.2) | <0.001 |
| Perc15 | 88.3 (19.6) | 65.9 (25.4) | <0.001 | 89.8 (19.0) | 0.413 |
| Pi10 | 2.1 (0.4) | 2.7 (0.6) | <0.001 | 1.9 (0.34) | <0.001 |
PRS Polygenic risk score, SGRQ St. George’s Respiratory Questionnaire, Perc15 CT attenuation at the 15th percentile of the lung CT histogram, Pi10 square root of the wall area of a theoretical airway with an internal perimeter of 10mm
We excluded 24 individuals who met the definition of both genetic and clinical resilience. Compared to genetic risk-matched individuals with COPD, genetically resilient individuals were younger, more likely to be African-American, more likely to be current individuals that smoke, had fewer pack-years, and less likely to have self-reported asthma. There was no difference in gender or BMI. Similarly, compared to low genetic risk individuals with COPD, genetically resilient individuals were younger and had fewer pack-years (Table S1). Compared to clinically resilient individuals, genetically resilient individuals were slightly younger, more likely to be African-American, had higher PRS, had higher BMI, and were less likely to have asthma. There were no differences in age, gender, current smoking status, or smoking pack-years. While there was no difference in peripheral vascular disease or comorbidity count, as defined by Putcha et al. [28], genetically resilient individuals were less likely to have atherosclerotic cardiovascular disease (stroke or heart attack) and had lower coronary artery calcium score compared to genetic risk-matched cases (Table S2). Conversely, there was no difference in atherosclerotic cardiovascular disease between genetically resilient and clinically resilient individuals.
Compared to both genetic risk-matched individuals with COPD and low genetic risk individuals with COPD, genetically resilient individuals without COPD had less disease severity as evidenced by higher FEV1 percent predicted, six-minute walk distance, lower SGRQ total score, lower BODE index, lower rate of bronchodilator response, and lower % LAA < −950 HU, lower percent gas trapping, higher Perc15, and lower Pi10. After additionally matching on age, race, and smoking pack-years, these differences persisted (Table S3). After adjustment for age, gender, race, current smoking status, smoking pack-years, BMI, and the first five principal components of genetic ancestry, the differences in FEV1 percent predicted between genetically resilient individuals and genetic risk-matched individuals with COPD remained significant (β = 35.9%, p < 0.001),
Compared to clinically resilient individuals, genetically resilient individuals demonstrated greater disease severity but similar amounts of emphysema, with a lower FEV1 percent predicted, lower six-minute walk distance, greater likelihood of having bronchodilator response, more gas trapping and higher Pi10, but no difference in % LAA < −950 HU or Perc15. After matching on age, race, and smoking pack-years, there was no difference in FEV1% predicted, 6MWD, bronchodilator response, but the differences in SGRQ, BODE index, gas trapping, and Pi10 persisted. In multivariable analysis, the difference between in FEV1 percent predicted between genetically resilient individuals and clinically resilient individuals was no longer significant (β = −1.5%, p = 0.144). Similarly, the difference in % LAA < −950 HU and Perc15 between genetically resilient individuals and risk-matched individuals with COPD persisted after adjustment for the covariates above as well as CT scanner model.
In sensitivity analyses comparing genetically resilient individuals who remain resilient at Phase 2 to clinically resilient individuals, genetically resilient individuals again had higher SGRQ, BODE index, and more gas trapping (Table S4). However, there was no difference in six-minute walk distance, FEV1 percent predicted, bronchodilator response, and Pi10. In non-Hispanic white individuals, the differences were similar between genetically resilient individuals and both genetic risk-matched individuals with COPD and clinically resilient individuals were similar to the larger population (Table S5). However, in African-American individuals, while the differences between genetically resilient individuals and genetic risk-matched individuals with COPD were similar, there was no difference in FEV1 percent predicted, % LAA < −950 HU, Perc15, or Pi10 (Table S6). After adjustment for age, gender, current smoking status, smoking pack-years, BMI, and ancestry-specific principal components, the differences in FEV1% predicted between genetically resilient individuals and both genetic risk-matched individuals with COPD (βNHW = 33.9%, p < 0.001; βAA = 35.4%, p < 0.001) and clinically resilient individuals (βNHW = −1.7%, p = 0.2; βAA = −1.4%, p = 0.4) were similar to the overall population.
Longitudinal outcomes across resilience subtypes and genetic risk-matched individuals with COPD
We examined the progression (i.e., change in COPD status) of individuals categorized into the resilience subtypes and genetic risk-matched individuals with COPD across the 5-year (Phase 2) and 10-year (Phase 3) visits. Among the genetic resilience group, 91 individuals (63.2%) remained resilient with 21 individuals (14.6%) progressing to COPD, 23 individuals (15.9%) progressing to PRISm, and 9 individuals (6.3%) had incomplete follow-up at Phase 2 (Fig. 2).
Fig. 2.
5-year and 10-year follow-up visit COPD status by resilience subtype. Alluvial diagram demonstrating flow of individuals from resilient subtype at baseline visit across 5-year and 10-year follow-up visit. COPD - chronic obstructive pulmonary disease; PRISm - preserved ratio impaired spirometry
At Phase 3, among individuals who had data through Phase 2, 20 individuals (13.9%) from the original genetically resilience group (n = 144) progressed to COPD, 14 individuals (9.7%) progressed to PRISm, 52 individuals (36.1%) remained resilient, 10 individuals (6.9%) had died, and 48 individuals (33.3%) had incomplete follow-up. The change in COPD status for the clinical resilience group and genetic risk-matched individuals with COPD are available in the supplement. Overall, excluding individuals who were lost to follow-up at Phase 3, while 71.6% of the clinically resilient group remained resilient at Phase 3, 54.2% of the genetically resilient group remained resilient at Phase 3.
We compared the change in clinical parameters from the baseline Phase 1 visit and the Phase 2 and 3 visits (Table 2).
Table 2.
Change clinical measures at 5-year and 10-year follow-up by resilience subtype
| Characteristic | Genetic Resilience | Genetic Risk-Matched Cases | p | Clinical Resilience | p |
|---|---|---|---|---|---|
| Change at 5-year follow-up visit (Phase 2) | |||||
| n | 135 | 302 | 420 | ||
| Change in FEV1 % predicted | -4.6 (9.0) | -3.9 (9.3) | 0.479 | 7.4 (5.9) | <0.001 |
| Change in FEV1, mL | -54.4 (46.4) | -39.4 (49.1) | 0.003 | 5.4 (23.7) | <0.001 |
| Change in SGRQ total score | 1.6 (16.8) | 0.8 (14.3) | 0.579 | 0.1 (10.4) | 0.203 |
| Change in six-minute walk distance, ft | -142.5 (337.2) | -154.0 (394.1) | 0.771 | -122.9 (306.1) | 0.532 |
| Change in percent emphysema | -0.5 (2.0) | 1.0 (5.3) | 0.002 | -0.2 (2.1) | 0.252 |
| Change at 10-year follow-up visit (Phase 3) | |||||
| n | 85 | 141 | 262 | ||
| Change in FEV1 % predicted | -6.2 (11.5) | -7.2 (9.6) | 0.515 | 4.1 (11.3) | <0.001 |
| Change in FEV1, mL | -46.2 (28.3) | -36.8 (25.3) | 0.01 | -24.0 (25.0) | <0.001 |
| Change in SGRQ total score | 4.2 (16.1) | 1.8 (17.4) | 0.293 | 2.2 (10.0) | 0.175 |
| Change in six-minute walk distance, ft | -252.2 (327.5) | -204.6 (418.3) | 0.4 | -202.0 (322.4) | 0.239 |
FEV1 Forced expiratory volume in one second, SGRQ St. George’s Respiratory Questionnaire
While there was no difference in FEV1 percent predicted, SGRQ total score, and six-minute walk distance, genetically resilient individuals showed a larger decrease in FEV1 in milliliters and a decrease in % LAA < −950 HU compared to genetic risk-matched cases. On the other hand, clinically resilient individuals had an increase in both FEV1 percent predicted and FEV1 in milliliters compared to genetically resilient individuals, but no difference in SGRQ total, six-minute walk distance, and % LAA < −950 HU. The pattern of change was similar from Phase 1 to Phase 3 across both comparisons.
We performed univariable and multivariable survival analyses across the three groups. A Kaplan-Meier curve is displayed in Fig. 3.
Fig. 3.
Survival probability after Phase 2 by resilience subtype. Kaplan-Meier curve demonstrating survival probability over time between genetic resilience, clinical resilience, and genetic risk-matched COPD. Time measured in days. COPD– chronic obstructive pulmonary disease
In a Cox proportional hazards model adjusted for age, race, gender, current smoking status, smoking pack-years, and BMI, when compared to individuals with COPD and high genetic risk, both clinical resilience (HR 0.34, 95% CI 0.23–0.50, p < 0.001) and genetic resilience (HR 0.41, 95% CI 0.23–0.73, p = 0.003) were associated with improved survival. In a separate Cox proportional hazards model directly comparing genetic and clinical resilience, there was no difference in survival (HR 0.78, 95% CI 0.41–1.47, p = 0.45). The proportional hazards assumptions were not violated in tested models.
Environmental and biological contributors to resilience
Given the observed differences in clinical, radiographic, and longitudinal characteristics between genetic and clinical resilience, we sought to identify possible etiologic differences. We found that genetically resilient individuals had higher (i.e., worse) SVI socioeconomic percentile, SVI minority status/language percentile, SVI housing type/transportation percentile, and SVI overall percentile compared to both genetic risk-matched individuals with COPD and clinically resilient individuals (Table 3).
Table 3.
Social determinants of health among COPD resilience subtypes
| Characteristic | Genetic Resilience | Genetic Risk-Matched Cases | p | Clinical Resilience | p |
|---|---|---|---|---|---|
| n | 138 | 323 | 419 | ||
| SVI Socioeconomic Percentile | 0.50 (0.30) | 0.42 (0.30) | 0.010 | 0.37 (0.28) | <0.001 |
| SVI Household Composition Percentile | 0.44 (0.28) | 0.45 (0.30) | 0.675 | 0.38 (0.29) | 0.038 |
| SVI Minority Status/Language Percentile | 0.60 (0.28) | 0.47 (0.27) | <0.001 | 0.47 (0.27) | <0.001 |
| SVI Housing Type/Transportation Percentile | 0.59 (0.28) | 0.51 (0.29) | 0.003 | 0.49 (0.30) | 0.001 |
| SVI Overall Percentile | 0.54 (0.29) | 0.44 (0.30) | 0.001 | 0.40 (0.29) | <0.001 |
| ADI National Rank | 40.60 (25.59) | 42.65 (27.62) | 0.461 | 40.97 (24.76) | 0.881 |
| ADI State Rank | 5.28 (2.82) | 4.93 (2.88) | 0.236 | 4.64 (2.85) | 0.023 |
In addition, SVI household composition percentile was higher in genetically resilient
SVI Social Vulnerability Index, ADI Area Deprivation Index
In addition, SVI household composition percentile was higher in genetically resilient individuals compared to clinically resilient individuals but not risk-matched individuals with COPD. Finally, the ADI state rank was higher for genetically resilient individuals compared to clinically resilient individuals but not genetic risk-matched individuals with COPD. We performed multivariable analyses, adjusted for age, gender, current smoking status, smoking pack-years, educational attainment, income category, insurance status, and ancestry-specific principal components, to test for the association of resilience subtype with SVI overall percentile after stratifying by race. While there was no difference in SVI overall percentile between African-American genetically resilient and clinically resilient individuals, the SVI overall percentile was significantly higher in non-Hispanic white genetically resilient individuals compared to non-Hispanic white clinically resilient individuals (βNHW = 0.09, p = 0.001). There was no difference in ADI state rank after stratification by race and adjustment for covariates between genetically resilient individuals and both comparison groups.
We examined protein expression between genetically resilient individuals and the two comparator groups, specifically selecting proteins that were previously examined in the clinical resilience study [29]. Compared to genetic risk-matched cases, there was no difference in CRP, Fibrinogen, tumor necrosis factor receptor super family member (TNFRSF) 1 A, and TNFRSF1B levels in blood in genetically resilient individuals (Table 4).
Table 4.
Expression of proteins associated with COPD across resilience subtypes
| Characteristic | Genetic Resilience | Genetic Risk-Matched Cases | p | Clinical Resilience | p |
|---|---|---|---|---|---|
| n | 144 | 362 | 420 | ||
| CRP | 14.3 (1.2) | 14.4 (1.2) | 0.425 | 13.9 (1.2) | 0.006 |
| Fibrinogen | 16.7 (0.6) | 16.7 (0.2) | 0.159 | 16.7 (0.2) | 0.4 |
| TNFRSF1A | 13.9 (0.5) | 13.9 (0.5) | 0.369 | 13.9 (0.4) | 0.46 |
| TNFRSF1B | 11.1 (0.4) | 11.1 (0.4) | 0.332 | 11.0 (0.3) | 0.011 |
CRP C reactive protein, TNFRSF1A Tumor necrosis factor receptor super family 1A, TNFRSF1B Tumor necrosis factor receptor super family 1B
Conversely, there were reduced CRP and TNFRSF1B levels in clinically resilient individuals compared to genetically resilient individuals.
Discussion
In this study, we applied a novel framework of genetic resilience to characterize individuals who are current or former individuals that smoke and have high genetic risk of COPD but appear resistant to the development of disease. We demonstrated that genetically resilient individuals have fewer symptoms, higher functional capacity, higher lung function, and fewer chest CT scan signs of disease compared to genetic risk-matched individuals with COPD. In addition, while genetically resilient individuals have similar clinical characteristics to clinically resilient individuals, there appears to be limited overlap in the number of genetically and clinically resilient individuals, which suggests that these two groups represent orthogonal subtypes of resilience in COPD. Finally, we found that, in the non-Hispanic white subpopulation, genetically resilient individuals had higher measures of social vulnerability compared to clinically resilient individuals. Taken together, our findings establish genetic resilience as a distinct phenotype of individuals at risk of developing COPD.
As expected, resilient individuals, by definition, did not show signs of COPD, including spirometry, functional status, and radiographic features. Thus, the baseline cross-sectional analyses reflect demographic differences may not to be attributable to genetic resilience. However, despite imbalanced demographic features, we were able to show that resilient individuals did not only appear resilient initially, but also that the majority of individuals demonstrated resilience to disease and mortality over time. Furthermore, by contrasting genetically resilient individuals to clinically resilient individuals, we were able to triangulate genetic resilience as a specific subtype beyond simply individuals unaffected by disease. However, given the number of individuals that met our definition relative to the overall study population, genetic resilience represents likely represents a small but nonetheless important subtype of individuals with smoking exposure.
While several of our findings confirmed our hypothesis regarding COPD-related differences between genetic risk-matched individuals with COPD, clinical resilience, and genetic resilience, there were several unexpected findings that may belie the differentiation of resilience subtypes in individuals at risk of COPD. First, our finding that genetically resilient non-Hispanic white individuals had higher social vulnerability compared to both genetic risk-matched individuals with COPD and clinically resilient individuals suggests that, in non-Hispanic white individuals: (1) at least some component of clinical resilience can be attributed to less social vulnerability; and (2) despite higher social vulnerability, genetically resilient individuals remain resilient to COPD. Rather than diminish clinical resilience as a relevant subtype in individuals at risk for COPD, this finding highlights the well-known contributions of the social determinants of health to the development, or in this case resistance to development, of disease [30–32]. Second, while we hypothesized that the two resilience subtypes would confer resistance to COPD and COPD-related measures of disease over time, we also found that both resilience subtypes were less likely to have a history of atherosclerotic cardiovascular disease as well as less coronary artery calcification measured on chest CT, both of which are common findings in individuals who smoke [33, 34]. This suggests that the mechanisms underlying both genetic resilience and clinical resilience provide resistance at least in part to the negative health effects of cigarette smoke beyond the lung given the shared pathobiological mechanisms underlying COPD and ASCVD [35].
Despite many strengths, we acknowledge that there are several limitations to our study. First, our sample size was limited due to the use of a single study population and a definition of high genetic risk in the highest decile. While we aim to expand our work to include other study populations in the future, we were still able to demonstrate important differences, which suggests a large true effect size of genetic resilience. In addition, while an empiric approach to defining the PRS threshold may yield a more precise result, prior studies examining genetic resilience in other complex diseases have successfully used a similar, a priori defined cutoff for high genetic risk [12, 36]. In addition, individuals in the top decile of PRS have been shown to have by far the highest risk of developing COPD. Thus, in balancing adequate sample size to detect clinically relevant differences while examining individuals at the highest risk, we feel our approach was rational, if not close to optimal. Second, by defining genetic resilience using lack of disease and a single, cross-sectional measure of lung function at the baseline visit, a higher proportion of individuals who were defined as genetically resilient went on to develop disease, as shown in Fig. 2, compared to clinically resilient individuals, who by definition had, at worst, “normal” lung function decline. Since the majority of large, population-based studies do not include longitudinal spirometric data, we opted to maintain the cross-sectional definition of genetic resilience, rather than incorporating a longitudinal component as is required by the SPIROMICS clinical resilience definition, in order to improve the generalizability of genetic resilience in COPD. In addition, given the complexity of the multi-dimensional definition of clinical resilience, a cross-sectional genomic definition may offer an attractive alternative. Third, we acknowledge that since the PRS was derived from a primarily European ancestry population and is less predictive of COPD in other genetic ancestries, including African American individuals. This may account for the higher proportion of African American individuals in the genetic resilience group. Nonetheless, many of the differences between the groups persisted after matching and adjustment for race. Finally, our insight into the biological mechanisms underlying resilience to COPD is limited for now. We hope to further explore the genetic basis of resilience by identifying genetic variants that ameliorate high genetic risk of COPD in the future.
In summary, our study reveals important clinical, radiographic, and longitudinal differences that separate genetically resilient individuals from similarly high-risk individuals with COPD as well as clinically resilient individuals. We show that genetic and clinical resilience have limited overlap and are largely distinct entities with potentially different underlying contributors, including environmental factors such as the social determinants of health. Further genetic and genomic studies will be required to elucidate the biological mechanisms that account for our observed findings and to translate them to novel therapies and interventions to slow and possibly prevent the development of COPD.
Supplementary Information
Acknowledgements
COPDGene was supported by NHLBI grants U01 HL089897 and U01 HL089856 and by NIH contract 75N92023D00011. COPDGene is also supported by the COPD Foundation through contributions made to an Industry Advisory Board comprised of AstraZeneca, Boehringer-Ingelheim, Genentech, GlaxoSmithKline, Novartis, Pfizer, Siemens and Sunovion.
Authors’ contributions
Study conceptualization: AJG, MM, SJG, CPH; Data acquisition: MHC, RPB, CPH; Data analysis: AJG, KYC, MM, JH, LPC, SP, SJG, CPH; Statistical support: AJG, KYC, MM, JH, MHC, SJG, CPH. All authors were responsible for the critical revision of the manuscript for important intellectual content.
Funding
AJG is supported by K08HL168205. MM is supported by K08HL159318. JLH is supported by R01NS128535 and the CNY Community Foundation. MHC is supported by R01HL168199, R01HL162813, and R01HL153248. SJG is supported by R01AG064955. CPH is supported by R01HL166231, R01HL168663, and K24HL173667.
Acknowledgements:
COPDGene was supported by NHLBI grants U01 HL089897 and U01 HL089856 and by NIH contract 75N92023D00011. COPDGene is also supported by the COPD Foundation through contributions made to an Industry Advisory Board comprised of AstraZeneca, Boehringer-Ingelheim, Genentech, GlaxoSmithKline, Novartis, Pfizer, Siemens and Sunovion.
Data availability
Data are available on the NCBI database of Genotypes and Phenotypes (dbGaP), accessions phs000179 and phs000765.
Declarations
Ethics approval and consent to participate
Institutional review boards approved the study at all participating centers. The current study was approved by the Mass General Brigham institutional review board (IRB #2007P000554) and SUNY Upstate Medical University institutional review board (IRB #2247707).
All participants provided written informed consent.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Auyon J. Ghosh, Ko-Yun Chang and Matthew Moll contributed equally and are co-first authors.
Stephen J. Glatt and Craig P. Hersh jointly supervised this study and are co-senior authors.
References
- 1.Venkatesan P. GOLD COPD report: 2024 update. Lancet Respir Med. 2024;12(1):15–6. [DOI] [PubMed] [Google Scholar]
- 2.Burrows B, Fletcher CM, Heard BE, Jones NL, Wootliff JS, THE EMPHYSEMATOUS AND BRONCHIAL TYPES OF CHRONIC AIRWAYS OBSTRUCTION. Lancet. 1966;287(7442):830–5. [DOI] [PubMed] [Google Scholar]
- 3.Kim V, Crapo J, Zhao H, et al. Comparison between an alternative and the classic definition of chronic bronchitis in COPDGene. Ann Am Thorac Soc. 2015;12(3):332–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Fletcher C, Peto R. The natural history of chronic airflow obstruction. BMJ. 1977;1(6077):1645–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Hersh CP, Make BJ, Lynch DA, et al. Non-emphysematous chronic obstructive pulmonary disease is associated with diabetes mellitus. BMC Pulm Med. 2014;14(1):164. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Castaldi PJ, Demeo DL, Hersh CP, et al. Impact of non-linear smoking effects on the identification of gene-by-smoking interactions in COPD genetics studies. Thorax. 2011;66(10):903–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.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):941–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Sakornsakolpat P, Prokopenko D, Lamontagne M, et al. Genetic landscape of chronic obstructive pulmonary disease identifies heterogeneous cell-type and phenotype associations. Nat Genet. 2019;51(3):494–505. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Hobbs BD, de Jong K, Lamontagne M, et al. Genetic loci associated with chronic obstructive pulmonary disease overlap with loci for lung function and pulmonary fibrosis. Nat Genet. 2017;49(3):426–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Moll M, Sakornsakolpat P, Shrine N, et al. 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] [PMC free article] [PubMed] [Google Scholar]
- 11.Chen R, Shi L, Hakenberg J, et al. Analysis of 589,306 genomes identifies individuals resilient to severe Mendelian childhood diseases. Nat Biotechnol. 2016;34(5):531–8. [DOI] [PubMed] [Google Scholar]
- 12.Hess JL, Tylee DS, Mattheisen M, et al. A polygenic resilience score moderates the genetic risk for schizophrenia. Mol Psychiatry. 2021;26(3):800–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Regan EA, Hokanson JE, Murphy JR, et al. Genetic epidemiology of COPD (COPDGene) study design. COPD: Journal of Chronic Obstructive Pulmonary Disease. 2011;7(1):32–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Bhatt SP, Balte PP, Schwartz JE, et al. Discriminative accuracy of FEV1:FVC thresholds for COPD-related hospitalization and mortality. JAMA. 2019;321(24):2438. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Hoffman EA, Ahmed FS, Baumhauer H, et al. Variation in the percent of emphysema-like lung in a healthy, nonsmoking multiethnic sample. The MESA lung study. Ann Am Thorac Soc. 2014;11(6):898–907. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Martinez CH, Diaz AA, Meldrum C, et al. Age and small airway imaging abnormalities in subjects with and without airflow obstruction in SPIROMICS. Am J Respir Crit Care Med. 2017;195(4):464–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Kohansal R, Martinez-Camblor P, Agustí A, Buist AS, Mannino DM, Soriano JB. The natural history of chronic airflow obstruction revisited. Am J Respir Crit Care Med. 2009;180(1):3–10. [DOI] [PubMed] [Google Scholar]
- 18.Cho MH, McDonald M-LN, Zhou X, et al. Risk loci for chronic obstructive pulmonary disease: a genome-wide association study and meta-analysis. Lancet Respir Med. 2014;2(3):214–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Kassambara A, Kosinski M, Biecek P, Fabian S. survminer: Drawing Survival Curves using ggplot2. 2019.
- 20.Therneau T, Lumley T. Survival Anal. 2019.
- 21.Kind AJH, Buckingham WR. Making neighborhood-Disadvantage metrics Accessible — The neighborhood atlas. N Engl J Med. 2018;378(26):2456–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Jiang R, Burke GL, Enright PL, et al. Inflammatory markers and longitudinal lung function decline in the elderly. Am J Epidemiol. 2008;168(6):602–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Ahmadi-Abhari S, Kaptoge S, Luben RN, Wareham NJ, Khaw K-T. Longitudinal association of C-reactive protein and lung function over 13 years: the EPIC-Norfolk study. Am J Epidemiol. 2014;179(1):48–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Garudadri S, Woodruff PG, Han MK, et al. Systemic markers of inflammation in smokers with symptoms despite preserved spirometry in SPIROMICS. Chest. 2019;155(5):908–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Dentener MA. Systemic anti-inflammatory mediators in COPD: increase in soluble interleukin 1 receptor II during treatment of exacerbations. Thorax. 2001;56(9):721–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Pratte KA, Curtis JL, Kechris K, et al. Soluble receptor for advanced glycation end products (sRAGE) as a biomarker of COPD. Respir Res. 2021;22(1):127. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Serban KA, Pratte KA, Strange C, et al. Unique and shared systemic biomarkers for emphysema in Alpha-1 antitrypsin deficiency and chronic obstructive pulmonary disease. EBioMedicine. 2022;84:104262. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Putcha N, Puhan MA, Drummond MB, et al. A simplified score to quantify comorbidity in COPD. PLoS ONE. 2014;9(12):e114438. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Oh AL, Mularski RA, Barjaktarevic I, et al. Defining resilience to smoking-related lung disease: a modified Delphi approach from SPIROMICS. Ann Am Thorac Soc. 2021;18(11):1822–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Lee Y-C, Chang K-Y, Mirsaeidi M. Association of county-degree social vulnerability with chronic respiratory disease mortality in the united States. Ann Am Thorac Soc. 2023;20(1):47–57. [DOI] [PubMed] [Google Scholar]
- 31.Hayes S, Duan KI, Wai TH, et al. Association between neighborhood socioeconomic disadvantage and chronic obstructive pulmonary disease prevalence among U.S. veterans. Ann Am Thorac Soc. 2024;21(4):669–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Carlson SA, Watson KB, Rockhill S, Wang Y, Pankowska MM, Greenlund KJ. Linking local-level chronic disease and social vulnerability measures to inform planning efforts: a COPD example. Prev Chronic Dis. 2023;20:230025. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Goff DC, Lloyd-Jones DM, Bennett G et al. 2013 ACC/AHA guideline on the assessment of cardiovascular risk. Circulation 2014;129(25_suppl_2). [DOI] [PubMed]
- 34.Oshunbade AA, Kassahun-Yimer W, Valle KA et al. Cigarette Smoking, incident coronary heart Disease, and coronary artery calcification in black adults: the Jackson heart study. J Am Heart Assoc. 2021;10(7):e017320 . [DOI] [PMC free article] [PubMed]
- 35.Rabe KF, Hurst JR, Suissa S. Cardiovascular disease and COPD: dangerous liaisons? Eur Respir Rev. 2018;27(149):180057. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Hou J, Hess JL, Armstrong N, et al. Polygenic resilience scores capture protective genetic effects for Alzheimer’s disease. Transl Psychiatry. 2022;12(1):296. [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
Data Availability Statement
Data are available on the NCBI database of Genotypes and Phenotypes (dbGaP), accessions phs000179 and phs000765.