Conceived and designed the experiments: MCA RK SBK BM YL EGB. Performed the experiments: MCA SBK BM CRG YL TBH EZ JZ MG TSL MF KL LJS EGB. Analyzed the data: MCA RK LAC LKW SS KL JG DH CRG YL MG MGF EGB. Contributed reagents/materials/analysis tools: YL. Wrote the paper: MCA RK EGB. Critical revision of the manuscript for important intellectual content: MCA RK LAC LKW SS SBK BM JG DH CRG YL TBH EZ JZ MG TSL MGF LJS MF KL EGB. Final approval of the version to be published: MCA RK LAC LKW SS SBK BM JG DH CRG YL TBH EZ JZ MG TSL MGF LJS MF KL EGB.
The authors have read the journal's policy and have the following conflicts: Dr. Aldrich has served as a Guest Editor for PLoS ONE. Christopher R. Gignoux holds stock with 23andMe, Inc. Dr. Smith received consultancy funds from Karmel-Sonix. Dr. Williams has received funding from Merck & Co for speaking engagements and travel. No other disclosures were reported. This does not alter the authors' adherence to all the PLoS ONE policies on sharing data and materials.
Smoking tobacco reduces lung function. African Americans have both lower lung function and decreased metabolism of tobacco smoke compared to European Americans. African ancestry is also associated with lower pulmonary function in African Americans. We aimed to determine whether African ancestry modifies the association between smoking and lung function and its rate of decline in African Americans.
We evaluated a prospective ongoing cohort of 1,281 African Americans participating in the Health, Aging, and Body Composition (Health ABC) Study initiated in 1997. We also examined an ongoing prospective cohort initiated in 1985 of 1,223 African Americans in the Coronary Artery Disease in Young Adults (CARDIA) Study. Pulmonary function and tobacco smoking exposure were measured at baseline and repeatedly over the follow-up period. Individual genetic ancestry proportions were estimated using ancestry informative markers selected to distinguish European and West African ancestry. African Americans with a high proportion of African ancestry had lower baseline forced expiratory volume in one second (FEV1) per pack-year of smoking (−5.7 ml FEV1/ smoking pack-year) compared with smokers with lower African ancestry (−4.6 ml in FEV1/ smoking pack-year) (interaction
African American individuals with a high proportion of African ancestry are at greater risk for losing lung function while smoking.
Tobacco smoking is the leading cause of preventable deaths in the United States and is an important contributor to accelerated lung function decline
Strong evidence from recent genome-wide association studies conducted in European American individuals implicates a genetic component to reduced lung function
All participants gave written informed consent. The Institutional Review Boards at the University of Pittsburgh, the University of Tennessee and the University of California, San Francisco approved the Health ABC protocol. The CARDIA study is reviewed annually by the Institutional Review Boards at the University of Alabama Birmingham, Northwestern University Feinberg School of Medicine, University of Minnesota and Kaiser Permanente in Oakland. CARDIA participants sign a new informed consent form at every examination.
Self-identified European American (N = 1,794) and African American participants (N = 1,281) were enrolled into the ongoing prospective Health, Aging and Body Composition (Health ABC) cohort study between April 1997 and June 1998. Subjects were eligible for participation if they were Medicare recipients, aged 70–79 years, and resided near Pittsburgh, Pennsylvania or Memphis, Tennessee. Additional eligibility criteria for the Health ABC Study participants included speaking English, being free from a life-threatening illness, and willingness to remain in the study area for ≥3 years. Persons were excluded if they met any of the following criteria: a) difficulty with daily living activities, defined as any difficulty walking a quarter of a mile or any difficulty walking up 10 steps without resting; b) cognitive impairment; c) inability to communicate with an interviewer; or d) treated for cancer within the previous three years. Further details of the Health ABC study population have been published elsewhere
Demographic data and detailed smoking histories were collected from Health ABC participants using a structured questionnaire during an extensive home interview conducted within year one of study enrollment. Smaller versions of the questionnaire were implemented during subsequent years of the study. Individuals who reported smoking <100 cigarettes were categorized as never smokers and individuals who reported smoking at least 100 cigarettes during their lifetime (i.e. five or more packs) were queried about smoking initiation, cessation and average cigarettes per day. Using this information we categorized participants as current smokers and former smokers and estimated the number of smoking pack-years (i.e., [no. cigarettes smoked per day*no. of years smoked]/20).
Methods employed for assessment of pulmonary function in Health ABC study participants are described elsewhere in greater detail
Health ABC participants provided a blood sample from which genomic DNA was extracted and whole genome amplified. A panel of 1,536 autosomal biallelic ancestry informative markers was selected for genotyping in Health ABC African Americans. Markers were selected to distinguish the continental ancestral populations comprising African Americans
The CARDIA Study is a prospective cohort initiated in 1985, with 5,115 participants enrolled between the ages of 18 to 30 years from four clinical sites. Of these, 2,637 self-identified as African American. Detailed study design, recruitment procedures, and baseline findings have been previously reported
Linear regression models were used to estimate the effect of African ancestry, coded as a continuous variable, on the association between tobacco smoke exposure and baseline FEV1 among African Americans without a self-reported diagnosis of pulmonary disease. Selection of covariates included in regression models was done using
Linear mixed effects models examined the influence of African ancestry on tobacco smoke exposure and FEV1 decline, adjusting for the same covariates as in baseline analyses. Restricted maximum likelihood estimation was implemented to examine the impact of African ancestry on the association between tobacco smoke exposure and FEV1 decline while simultaneously accounting for the correlation structure between repeated pulmonary measures over the 10-year follow-up period in Health ABC. Smoking status was allowed to vary over time and continuous measures were mean centered. Similar linear mixed effects models were evaluated in CARDIA examining the relationship between African ancestry and smoking status (never, current, and former) on rate of pulmonary function decline. Covariates included sex, clinic, height, height-squared, body mass index (BMI), smoking pack-years and maximum attained education. Three-way interaction terms between age (as a measure of time), percent African ancestry, and smoking status were included in mixed models to estimate p-values for interactions. A two-sided significance probability of 0.05 was used to infer presence or absence of interaction. A p-value of 0.2 indicated suggestive evidence for the presence or absence of interaction, as a recommended strategy to reduce type II error and avoid missing a true interaction
In its initial year, 1,281 African Americans, mean age 73.9 years, were enrolled in the Health ABC study. Repeat measures of pulmonary function were obtained during years one, five, eight and ten. At year one 942 African Americans had acceptable spirometry results and during years five, eight and ten there were 587, 406, and 337 individuals, respectively, had acceptable spirometry results; the remaining study individuals were either lost to follow-up or deceased over the ten year follow-up period (see
Characteristic | N (%) | |
Age, mean years (SD) | 73.9 (2.9) | |
Male sex (%) | 552 (43.0) | |
Education, years (%) | ||
< High school | 556 (43.6) | |
High school graduate | 389 (30.5) | |
> High school | 330 (25.9) | |
Unknown | 6 | |
Income, $ (%) | ||
<10,000 | 296 (26.2) | |
10,000–25,000 | 553 (48.9) | |
25,001–49,999 | 225 (19.9) | |
≥ 50,000 | 56 (5.0) | |
Unknown | 151 | |
Smoking status (%) | ||
Never smokers | 573 (44.8) | |
Current smokers | 207 (16.2) | |
Former smokers | 498 (39.0) | |
Unknown | 3 | |
Smoking pack-years |
30.3 (25.4) | |
Age initiated smoking, years, mean (SD) | 19.6 (7.0) | |
Age stopped smoking, years, mean (SD) | 51.0 (14.0) | |
Average cigarettes smoked per day |
12.4 (8.6) | |
Pulmonary function, mean (SD) | ||
FEV1 (milliliters) | 1972.2 (575.7) | |
FVC (milliliters) | 2551.8 (724.9) | |
% African ancestry, mean (SD) | 78.5 (13.1) | |
Unknown | 126 |
Definition of abbreviations: BMI = body mass index; FEV1 = forced expiratory volume at one second; FVC = forced vital capacity.
Categorical and continuous variables were assessed with chi-square test and two-sample t-test, respectively.
among ever smokers only.
among current smokers only.
A total of 813 African Americans had both acceptable quality spirometry data and individual genetic ancestry estimates. Multivariable linear regression analyses demonstrated an inverse relationship between FEV1 and both African ancestry and amount smoked. Smoking was associated with a lower FEV1 among individuals with both high African ancestry (defined as greater than median value of 80.8%) and those with low African ancestry (less than or equal to 80.8%). The reduction in baseline FEV1 associated with smoking was greater among those with higher African ancestry.
<80.8% African ancestry | ≥80.8% African ancestry | |||||
(N = 169) | (N = 196) | |||||
Variable | Beta | SE | Beta | SE | ||
Intercept | 2121.67 | 165.37 | <0.001 | 2126.33 | 121.53 | <0.001 |
Age, years | −14.36 | 12.16 | 0.24 | −14.56 | 9.66 | 0.13 |
Female | −383.46 | 100.53 | <0.001 | −375.70 | 80.60 | <0.001 |
Smoking, pack-years | −4.58 | 1.50 | 0.003 | −5.69 | 1.21 | <0.001 |
Standing height, mm | 18.72 | 5.18 | <0.001 | 21.29 | 4.32 | <0.001 |
Household income | 57.59 | 41.77 | 0.17 | 9.71 | 36.71 | 0.79 |
Pittsburgh | 47.11 | 72.90 | 0.52 | 64.66 | 57.87 | 0.27 |
Definition of abbreviations: SE = standard error; mm = millimeter.
Observations with standardized residuals > 2.5 were removed. Median African ancestry is 80.8%, estimated from 1332 ancestry informative markers.
Never smokers | Ever smokers | |||||
(N = 270) | (N = 371) | |||||
Variable | Beta | SE | Beta | SE | ||
Intercept | 2081.05 | 212.02 | <0.001 | 2354.20 | 190.92 | <0.001 |
% African ancestry |
−1.43 | 1.92 | 0.46 | −4.40 | 1.80 | 0.02 |
Age, years | −19.62 | 7.85 | 0.01 | −13.44 | 7.75 | 0.08 |
Female | −233.20 | 63.96 | <0.001 | −340.92 | 63.98 | <0.001 |
Standing height, mm | 24.12 | 3.51 | <0.001 | 19.21 | 3.37 | <0.001 |
Household income | 35.19 | 30.49 | 0.25 | 30.46 | 28.21 | 0.28 |
Pittsburgh | 95.33 | 48.29 | 0.05 | 30.62 | 46.15 | 0.51 |
Definition of abbreviations: SE = standard error; mm = millimeter.
Observations with standardized residuals >2.5 were removed.
African ancestry is coded as a continuous variable.
Multivariable linear mixed effects models examining the rate of decline in FEV1 revealed a three-way interaction suggestive of a complex relationship between age, smoking, and individual African ancestry in the Health ABC study population, although results did not meet statistical significance (
Health ABC (A) and CARDIA (B). Current smokers are represented by the ▴ symbol and former smokers by the • symbol. Low African ancestry groups are represented by a solid line and high African ancestry by a dashed line. The groups are as follows: low African current smokers (solid line with ▴ symbol), high African current smokers (dashed line with ▴ symbol), low African former smokers (solid line with • symbol), and high African former smokers (dashed line with • symbol).
FEV1 (ml/yr) | |||
(N = 1,737) | |||
Variable | Beta | SE | |
|
|||
Intercept | 1630.22 | 58.43 | <0.001 |
African ancestry, % | −2.90 | 1.40 | 0.04 |
Age, year at time t | −26.45 | 10.18 | 0.01 |
Smoking, pack-years | −4.60 | 0.79 | <0.001 |
Age*African*Pack-years | 0.007 | 0.004 | 0.14 |
|
|||
Intercept | 1629.17 | 58.66 | <0.001 |
African ancestry, % | −4.37 | 1.70 | 0.01 |
Age, year at time t | −31.55 | 2.24 | <0.001 |
Smoking, pack-years | −4.61 | 0.77 | <0.001 |
Age*African*Former smoker | 0.32 | 0.23 | 0.16 |
Age*African*Current smoker | 0.12 | 0.43 | 0.78 |
Mixed effects models are adjusted for sex, standing height, household income, clinic site and two-way interactions among covariates. Age, height and African ancestry are mean centered.
Model 1 is also adjusted for smoking status.
Using two independent adult cohorts, we demonstrated that smoking has a greater impact on reduced lung function among African Americans with high African ancestry than those with low African ancestry. Importantly, we found smokers with high African ancestry had lower baseline lung function than individuals with lower African ancestry. Inclusion of the 1,332 AIMs separately in models examining the association between tobacco smoke exposure and baseline FEV1 did not abrogate the response, even after correcting for multiple testing using false discovery rate
Prior to our study, other groups have found inconsistent associations between race and smoking associated lung function decline. Vollmer
Our findings suggest that tobacco-related lung damage and African ancestry together have a substantially greater impact on lung function and perhaps its decline than either factor alone. In other words, the relationship between smoking and pulmonary function depends on the level of African ancestry. Individuals with high African ancestry are particularly susceptible to the impact of cigarette smoking on the decline of FEV1. In contrast to current smokers, former smokers with high African ancestry have lung function that becomes more similar to individuals with low African ancestry, suggesting lung function improves with smoking cessation irrespective of African ancestry. In fact, former smokers with high African ancestry may have an even greater benefit from smoking cessation, such that their lung function approaches that of former smokers with low African ancestry. These findings reveal greater complexity in the relationship between race and tobacco smoking associated lung function decline than was seen in an earlier study
To assess potential survival bias, we compared baseline characteristics between those lost to follow-up and individuals remaining in the Health ABC study. Individuals remaining in the study were more likely to be male, higher BMI, higher education, fewer smoking pack-years, and lower African ancestry (
Health ABC provides a robust approach for observing baseline spirometry measures, but also allows for assessment of longitudinal changes in lung function associated with smoking in this unique population of older African American adults. Replication of our findings in the CARDIA African American participants provides additional support. Few studies have ascertained repeated pulmonary function measures in either older populations or African American individuals. Thus the collection of repeated pulmonary function measures over a ten year time period in Health ABC and over 20 years in CARDIA is a notable strength of this study. Additionally, stringent protocols were implemented to ensure that quality spirometry measures were obtained. The prospective design of CARDIA and Health ABC minimizes possible participant response bias in reporting of smoking exposures.
In summary, we have demonstrated that African ancestry and tobacco use act synergistically to accelerate rate of lung function decline. Specifically, our results suggest that individuals with high African ancestry have increased risk of smoking-related changes in lung function and may be an important group for smoking intervention. In addition, the observation that former smokers with high African ancestry may have a less rapid rate of decline in lung function suggests that this group may preferentially benefit from cessation. Our findings have important public health implications. We have demonstrated that genetic ancestry may serve as a biomarker for identifying smokers who would benefit from targeted counseling regarding smoking cessation
(PDF)
(PDF)
(PDF)
(PDF)
(PDF)
(PDF)