Author + information
- Received May 29, 2018
- Accepted June 18, 2018
- Published online August 27, 2018.
- Odilson M. Silvestre, MD, MPHa,b,
- Wilson Nadruz Jr., MDa,c,
- Gabriela Querejeta Roca, MDa,
- Brian Claggett, PhDa,
- Scott D. Solomon, MDa,
- Maria C. Mirabelli, MDd,
- Stephanie J. London, MDe,
- Laura R. Loehr, MDf and
- Amil M. Shah, MD, MPHa,∗ (, )@BrighamWomens
- aDivision of Cardiovascular Medicine, Brigham and Women’s Hospital, Boston, Massachusetts
- bDepartment of Internal Medicine, Federal University of Acre, Rio Branco, Acre, Brazil
- cDepartment of Internal Medicine, University of Campinas, Campinas, Brazil
- dRollins School of Public Health, Emory University, Atlanta, Georgia
- eNational Institute of Environmental Health Sciences, National Institutes of Health, Department of Health and Human Services, Research Triangle Park, North Carolina
- fGillings School of Public Health, University of North Carolina, Chapel Hill, North Carolina
- ↵∗Address for correspondence:
Dr. Amil M. Shah, Division of Cardiovascular Medicine, Brigham and Women’s Hospital, 75 Francis Street, Boston, Massachusetts 02115.
Background Pulmonary dysfunction predicts incident cardiovascular disease (CVD).
Objectives The purpose of this study was to evaluate whether longitudinal decline in lung function is associated with incident heart failure (HF), coronary heart disease (CHD), and stroke.
Methods Among 10,351 participants in the ARIC (Atherosclerosis Risk In Communities) study free of CVD, rapid lung function decline was defined as the greatest quartile (n = 2,585) of decline in either forced expiratory volume in 1 s (FEV1) (>1.9% decline/year) or forced vital capacity (FVC) (>2.1% decline/year) over 2.9 ± 0.2 years. The relationship between rapid decline in FEV1 or FVC and subsequent incident HF, CHD, stroke, or a composite of these was assessed using multivariable Cox regression adjusting for the baseline spirometry value, demographics, height, body mass index, heart rate, diabetes, hypertension, low-density lipoprotein, use of lipid-lowering medication, N-terminal fragment of prohormone for B-type natriuretic peptide, and smoking.
Results The mean age was 54 ± 6 years, 56% were women, and 81% were white. At 17 ± 6 years of follow-up, HF occurred in 14%, CHD 11%, stroke 6%, and the composite in 24%. Rapid decline in FEV1 and in FVC were both associated with a heightened risk of incident HF (hazard ratio [HR]: 1.17; 95% confidence interval [CI]: 1.04 to 1.33; p = 0.010; and HR: 1.27; 95% CI: 1.12 to 1.44; p < 0.001; respectively), with rapid decline in FEV1 most prognostic in the first year of follow-up (HR: 4.22; 95% CI: 1.34 to 13.26; p = 0.01). Rapid decline in FEV1 was also associated with incident stroke (HR: 1.25; 95% CI: 1.04 to 1.50; p = 0.015).
Conclusions A rapid decline in lung function, assessed by serial spirometry, is associated with a higher incidence of subsequent CVD, particularly incident HF.
Cardiovascular disease (CVD) is the leading cause of death in the United States, while chronic pulmonary disease is the third leading cause of mortality (1). Interactions between cardiac and pulmonary function in relation to clinical outcomes are well recognized (2). Impaired lung function is associated with nearly 2-fold higher risk of cardiovascular mortality (3). Conversely, coronary heart disease (CHD) and heart failure (HF) are more common among people with pulmonary disease, even after accounting for common cardiovascular risk factors, including smoking (4,5). Asthma and chronic obstructive pulmonary disease (COPD), the most prevalent respiratory diseases worldwide (6), are associated with a systemic inflammatory response (7) and with systemic endothelial dysfunction and impaired vascular reactivity (8), which also promote CVD (9) and represent potential mechanisms underlying these cardiopulmonary interactions.
Age-related decline in lung function starts at approximately 25 years of age, with an average loss of 20 ml/year in forced expiratory volume in 1 s (FEV1) (10). Forced vital capacity (FVC) also declines with age to approximately 75% of the best value (10). Environmental and behavioral factors, most notably cigarette smoking, in addition to genetics, interact to determine the speed of the decline in the lung function (11). A rapid decline in lung function is a predictor of incident COPD (12,13). However, although rate of decline in lung function is associated with coronary disease mortality (14,15), the association with incident cardiovascular events is not known. Furthermore, incipient or early HF may cause rapid deterioration in spirometric measures, and FEV1 in particular, due to interstitial and alveolar edema and consequent airway compression (16). However, rapid lung function decline secondary to early and undiagnosed HF (reserve causality) would be expected to predict incident HF during short-term—as opposed to long-term—follow-up. We hypothesize that a greater decline in pulmonary function will be associated with a heightened risk of incident HF, stroke, coronary disease, and death over 20-year follow-up among middle-aged persons from a community-based sample free of prevalent CVD.
ARIC (Atherosclerosis Risk In Communities) is a prospective cohort study of 15,792 persons, between ages 45 and 64 years at the time of recruitment (17). Participants were recruited using a probability sampling approach in 4 different centers in the United States: Forsyth County, North Carolina; Jackson, Mississippi; the northwest suburbs of Minneapolis, Minnesota; and Washington County, Maryland. Since study inception (1987 to 1989, visit 1), 4 subsequent study visits were performed. Spirometry was performed at visits 1 and 2 (1990 to 1992) in all participants attending each visit. The study was approved by institutional review boards from each site, and all participants gave written informed consent.
For this analysis, we included participants who underwent spirometry as part of visit 1 (1987 to 1989) and visit 2 (1990 to 1992). We excluded subjects with prevalent HF (n = 700), CHD (n = 661), and stroke (n = 177) at visit 2; those without information on HF (n = 234) and death (n = 1); and participants without spirometry data at visits 1 or 2 (n = 312) (Online Figure 1). We additionally excluded subjects with poor performance on spirometry as indicated by tests lasting <6 s or with irregular tracing, or with spirograms that were considered not reproducible (n = 1,912) following the previously published American Thoracic Society guidelines (18). A total of 10,351 participants were included in this analysis.
Assessment of lung function and rapid decline in lung function
Spirometry was performed following the American Thoracic Society guidelines at visits 1 and 2 (18). Spirometry was conducted using a water-sealed Collins Survey II volume displacement spirometer (Collins Medical, Fairfield, Connecticut) and Pulmo-Screen II software (PDS Healthcare Products, Brookfield, Wisconsin). Three or more acceptable spirograms were obtained from at least 5 forced expirations. The best single spirogram was identified by a computer and confirmed by a technician (18). FEV1 was assessed as the volume of gas exhaled in the first second of expiration. FVC was assessed as the volume of gas forcefully exhaled after maximal inspiratory effort. Equations for predicting FEV1 and FVC were developed to obtain the percentage of predicted FEV1 and FVC adjusted for age, sex, race, and height in the ARIC study (19). The primary exposure variable for this analysis was the change in percent predicted FEV1 or change in percent predicted FVC between visits 1 and 2 (visit 1 value − visit 2 value for both). We primarily used FEV1 and FVC as percentage of predicted instead of absolute values because the percentage predicted is a more reliable measurement of changes in spirometry over a time period <5 years (20). Supplemental analyses were performed with the FEV1/FVC ratio as the primary predictor.
Ascertainment of incident cardiovascular events
Incident HF, CHD, stroke, and death were ascertained through 2012 from ongoing surveillance methods as previously described (21,22). Incident HF was defined as the first occurrence of a hospitalization with an HF diagnosis according to the International Classification of Diseases-9th Revision (ICD-9) code 428 (428.0 to 428.9) in any position ascertained by the ARIC study retrospective surveillance of hospital discharges, or a death certificate with death from HF in any position or with an ICD-9 code of 428 or an ICD-10 code of I50 among any of the listed diagnoses or underlying causes of death (23). CHD was defined as a hospitalized acute myocardial infarction or CHD mortality. CHD was ascertained from yearly follow-up phone calls, study visits, hospital discharge information on fatal and nonfatal myocardial infarction, next-of-kin interviews, physician-completed questionnaires, and death certification, and was adjudicated by the ARIC Mortality and Morbidity Classification Committee as previously described (21). Stroke was defined as rapid onset of a focal neurological deficit lasting >24 h or until death in the absence of a nonstroke cause. A stroke event was identified from annual phone calls, visits, or surveillance of ARIC community hospitals. A hospitalization was considered for validation of stroke event if the discharge diagnosis contained a cerebrovascular disease diagnosis code (ICD-9 codes 430 to 438), if the discharge summary included a key word related to cerebrovascular procedure or disease, or if there was imaging evidence of cerebrovascular disease. Following the National Survey of Stroke criteria for stroke definition (24), a computerized algorithm and a physician reviewer independently classified strokes as definitive or probable (22). Death was ascertained through linkage with the National Death Index.
We used covariates measured at both visits 1 and 2. Race was self-reported. Hypertension was defined as systolic blood pressure ≥140 mm Hg, diastolic blood pressure ≥90 mm Hg, or use of antihypertensive medication. Body mass index (BMI) was calculated by dividing measured weight (in kilograms) by height (in meters) squared and was defined as previously described in the ARIC study (25). Diabetes mellitus was defined as a baseline fasting glucose level ≥126 mg/dl, baseline nonfasting glucose level ≥200 mg/dl, self-reported physician diagnosis of diabetes, or the presence of medication for diabetes. Smoking history was ascertained by an interviewer-administered questionnaire, and categorized as never smokers, former smokers, and current smokers. Pack-years were obtained from the multiplication of average number of cigarettes/day and number of years of smoking reported until visit 2 as previously described (26,27). The N-terminal fragment of prohormone for B-type natriuretic peptide (NT-proBNP) and the high-sensitivity C-reactive protein (hs-CRP) were measured from visit 2 serum samples that had previously been stored at −70°C. NT-proBNP measurements were performed with a sandwich immunoassay method on the Roche Elecsys 2010 Analyzer (Roche Diagnostics, Indianapolis, Indiana). The assay coefficients of variation were 5.4% at an NT-proBNP value of 133 pg/ml and 4.3% at a value of 4,516 pg/ml. The hs-CRP was measured using an immunoturbidimetric assay on the Roche Modular P chemistry analyzer (Roche Diagnostics). The coefficient of variation was 7.0% (28).
For each spirometry measure, initial quartile analysis suggested nonlinear associations of change between visits 1 and 2 and cardiovascular outcomes, with risk primarily associated with the greatest quartile of decline (Online Table 1). Therefore, change in each spirometric measure was dichotomized, with rapid decline defined as the greatest quartile of change between visits 1 and 2. Nonrapid decliners were defined as those participants in the remaining 3 quartiles and served as the reference group. This approach is concordant with several prior publications assessing rapid decline in lung function (14,29). Participant characteristics were described among rapid and nonrapid decliners. Categorical data were reported as percent frequencies and compared by chi-square test. Continuous normally distributed data were displayed as mean ± SD, and continuous non-normally distributed data were displayed as median and 25th and 75th percentile. Comparisons of continuous variables between rapid decliners and nonrapid decliners were performed using the Student’s t-test or Wilcoxon rank sum test, accordingly.
Rates of incident HF, CHD, stroke, death, and the composite of these are presented as events per 1,000 person-years at risk. The risk associated with rapid decline was assessed using multivariable Cox proportional hazards regression models with the nonrapid decliner group as the reference population. Three models were constructed adjusting for covariates selected based on prior literature. Model 1 was unadjusted; model 2 adjusted for the baseline spirometry value, age, sex, race, and height; and model 3 additionally adjusted for ARIC field center, BMI, heart rate, low-density lipoprotein cholesterol, use of lipid-lowering medications, NT-proBNP, diabetes, hypertension, and smoking (smoking status and pack-years). Outcomes included HF, CHD, and stroke individually, the composite of these (HF, CHD, and stroke), and the composite of these events in addition to death (HF, CHD, stroke, and death). Effect modification of age, race, and baseline spirometry on the relationship between rapid decline in lung function and cardiovascular outcomes was assessed using multiplicative interaction terms. The proportional hazards assumption was tested for all models. In analyses where a violation of the proportional hazards assumption was detected with the plots of log (−log [survival function]) vs. log (time) for the association between rapid decline in FEV1 and HF, we also assessed the relationship over more restricted time intervals, partitioning the follow-up time as less than or more than 1 year, and as less than or more than 10 years. These analyses were performed by placing landmarks at 1 and 10 years, respectively, and removing all patients who had a previous event or had died. Second events were not taken into account. We assessed for competing risk of death for cardiovascular events using the Fine and Gray competing risks regression model to estimate the corresponding subdistribution hazard ratio (30). All statistical analyses were performed with STATA version 13.1 (Stata Corp., College Station, Texas). All p values <0.05 were considered statistically significant.
The mean age of the cohort was 53.7 ± 5.7 years, 56% were women, and 81% were white (Tables 1 and 2⇓⇓). By definition, 25% of participants were classified as rapid decliners by the FEV1 criteria of a >1.9%/year decrease in FEV1, and 25% by the FVC criteria of a >2.1%/year decrease in FVC. A total of 16% of participants met the definition of rapid decline by both FEV1 and FVC criteria. Compared to ARIC participants included in this analysis, those not included due to poor quality or missing spirometry at visit 2 tended to be older and more frequently black, with a higher BMI and higher prevalence of hypertension, diabetes, and smoking (Online Table 2).
Rapid decline in FEV1 and incident cardiovascular events
The mean annual rate of decline in percent predicted FEV1 was 3.27 ± 1.79%/year for FEV1 rapid decliners and 0.02 ± 1.75%/year for nonrapid decliners. Rapid decliners started with a higher FEV1 at visit 1 but ended with a lower FEV1 at visit 2 (Table 1). FEV1 rapid decliners also demonstrated a greater annual decline in percent predicted FVC compared with nonrapid decliners (3.09 ± 2.63%/year vs. 0.37 ± 1.81%/year, respectively).
Compared with nonrapid decliners, FEV1 rapid decliners were older, more likely to be male, slightly taller, and had a higher prevalence of current smoking (Table 1). The prevalence of diabetes and hypertension were similar between groups, as were cholesterol levels. At visit 1, BMI and blood pressure were not different among rapid and nonrapid decliners. However, by visit 2, rapid decliners had higher BMI and blood pressure levels. High-sensitivity CRP, a marker of systemic inflammation, was significantly higher among rapid decliners (2.4 mg/l [25th–75th percentiles: 1.2–5.0 mg/l] vs. 2.0 mg/l [25th–75th percentiles: 0.9–4.1 mg/l]; p < 0.001). NT-proBNP assessed at visit 2 was also higher among rapid decliners (53.8 pg/ml [25th–75th percentiles: 27.9–99.7 pg/ml] vs. 47.8 pg/ml [25th–75th percentiles: 25.5–84.7 pg/ml]; p < 0.001).
At a mean follow-up of 17 ± 6 years, CVD occurred in 2,490 (24%) participants. HF occurred in 1,473 (14%), CHD in 1,093 (11%), stroke in 664 (6%), and death in 2,852 (28%) (Table 3). These incidence rates are similar to those reported in other general U.S. population studies (31–34). FEV1 rapid decliners showed a higher rate of incident CVD compared with nonrapid decliners (composite of HF, CHD, and stroke: 16.7 [95% confidence interval (CI): 15.5 to 18.0] per 1,000 person-years and 12.9 [95% CI: 12.3 to 13.5] per 1,000 person-years, respectively) (Figure 1). After adjusting for baseline FEV1, age, sex, race, height, ARIC center, BMI, heart rate, low-density lipoprotein cholesterol, NT-proBNP, diabetes, hypertension, and smoking, rapid FEV1 decline remained associated with a higher risk of the composite endpoint (hazard ratio [HR]: 1.15; 95% CI: 1.04 to 1.26; p = 0.004) (Table 3, Central Illustration). Rapid decliners also had a higher risk of death (Table 3). After accounting for competing risk of noncardiovascular death, the risk of incident CVD associated with rapid decline in FEV1 remained similar to that observed in the primary analysis (Table 3).
Rapid decliners had a higher risk of incident HF (HR: 1.17; 95% CI: 1.04 to 1.33; p = 0.010), which persisted after excluding the 1,093 subjects with incident CHD (HR: 1.21; 95% CI: 1.03 to 1.39; p = 0.013). A violation of the proportional hazard assumption (p = 0.003) was observed for the relationship between decline in FEV1 and incident HF, such that the risk associated with FEV1 rapid decline was greater during the early follow-up period and attenuated during late follow-up. For example, FEV1 rapid decline was associated with a higher risk of HF in the first 1 year of follow-up (Table 4, Figure 1, Online Figure 2). Similarly, rapid decline was associated with a higher risk of HF during the first 10 years of follow-up, but not after this point (Table 4). Rapid decline in FEV1 was also associated with a heightened risk of stroke (HR: 1.25; 95% CI: 1.04 to 1.50; p = 0.015) (Table 3), with no evidence of a change in the magnitude of risk over the follow-up period. Rapid decline in FEV1 was not associated with incident coronary disease (Table 3). In supplemental analyses using rapid decline in the FEV1/FVC ratio, rapid decline in the ratio was similarly associated with incident HF and with the composite of HF, CHD, and stroke (Online Table 3).
Rapid decline in FVC and incident cardiovascular events
FVC rapid decliners had a lower FVC at visit 1 compared with nonrapid decliners (Table 2), and demonstrated an annual decline of 3.70 ± 2.31%/year compared with 0.17 ± 1.60%/year for nonrapid decliners. FEV1 was similar between these groups at baseline, but demonstrated greater decline in the FVC rapid compared with nonrapid decline groups (2.69 ± 2.18%/year vs. 0.17 ± 1.91%/year, respectively) (Table 2). FVC rapid decliners were older and more likely to be black. At visit 1, they had higher BMI, heart rate, prevalence of hypertension, and active smoking (Table 2). These patterns persisted at visit 2, but the magnitude of the difference in BMI, heart rate, and blood pressure between groups was more prominent. Both hs-CRP and NT-proBNP were higher among rapid decliners (2.6 mg/l [25th–75th percentiles: 1.2–5.4 mg/l] vs. 1.9 mg/l [25th–75th percentiles: 0.9–4.0 mg/l]; p < 0.001; and 53.4 pg/ml [25th–75th percentiles: 27.6–98.9 pg/ml] vs. 47.7 pg/ml [25th–75th percentiles: 25.7–85.3 pg/ml]; p < 0.001, respectively).
FVC rapid decline was associated with a higher rate of the composite of incident HF, CHD, and stroke (incidence rate: 16.6 [95% CI: 15.5 to 18.0] per 1,000 person-years in rapid decliners and 12.6 [95% CI: 12.0 to 13.2] per 1,000 person-years in nonrapid decliners) (Table 5, Figure 2). After full adjustment, rapid FVC decline was associated with a 19% higher risk of composite endpoint (HR: 1.19; 95% CI: 1.08 to 1.32; p < 0.001) (Table 5). FVC rapid decline was also associated with a heightened risk of death (Table 5). In competing risk models, death was not a significant competing risk for CVD. For components of this composite, FVC rapid decline was associated with an increased risk of incident HF (HR: 1.27; 95% CI: 1.12 to 1.44; p < 0.001), which persisted after excluding participants with incident CHD (HR: 1.29; 95% CI: 1.11 to 1.50; p = 0.001). Unlike FEV1 rapid decline, no significant variation in the risk associated with FVC rapid decline was noted through the follow-up period. No association was noted between FVC rapid decline and incident stroke or CHD (Table 4, Figure 2).
No heterogeneity of effect by race or sex was detected for the association between either rapid decline in FEV1 or in FVC with HF, CHD, or stroke (all p for interaction >0.05). Baseline FEV1 did significantly modify the relationship between rapid decline in FEV1 and incident HF (p for interaction 0.004), such that the association of rapid decline in FEV1 was only observed among participants with a percent predicted FEV1 <80% at baseline (Online Table 4). Baseline FEV1 did not significantly modify the relationship between rapid decline in FEV1 and the other outcomes assessed. Similarly, no interaction was noted between baseline FVC and the relationship between rapid decline in FVC and any of the studied outcomes. Similar associations with outcomes were noted when change in both percent predicted FEV1 and FVC were modeled as continuous variables instead of categorical variables (Online Table 5). In a sensitivity analysis stratified by smoking status (current and former smokers vs. never smokers), similar findings were noted as in the primary analysis (Online Tables 6 and 7). No heterogeneity of effect was noted, except for the relationship between rapid decline in FEV1 and incident CHD (p for interaction by smoking status = 0.04), such that rapid decline in FEV1 predicted incident CHD among never smokers but not smokers. Similar findings as in the primary analysis were also noted after exclusion of participants with probable COPD at visit 1 (2.5% of participants) (Online Table 8).
In a biracial community-based sample of 10,351 participants followed for approximately 17 years, rapid decline in lung function is associated with a higher risk of CVD, independent of established cardiovascular risk factors (Central Illustration). Rapid decline in FEV1 is associated with a higher incidence of HF, stroke, and death, while rapid decline in FVC is associated with a higher incidence of HF and death. Notably, decline in FEV1 was predictive of incident HF mainly in the first year of follow-up. In contrast, a consistent risk throughout the follow-up period was noted for the association of decline in FVC with incident HF, and for the association of both exposure measures with the other study endpoints. Neither sex nor race significantly modified these associations. These findings demonstrate that deterioration in lung function is a predictor of incident CVD, independent of smoking status and baseline lung function.
Decline in lung function starts in young adulthood. An annual fall of 25 to 30 ml in FEV1 is considered normal (13,35). Accelerated decline in pulmonary function results from a combination of genetic determinants and environmental exposure, and increases the risk of pulmonary diseases, mainly COPD (36). Previous studies have shown an association between accelerated decline in lung function with systemic inflammatory activation (37), decline in arterial elasticity (38), and hypertension (39). Importantly, these same pathways have also been implicated in the development of atherosclerotic CHD (40,41), ischemic stroke (42), and HF (43). Recent attention has focused on the role of comorbidity-driven systemic inflammation in the development of LV diastolic dysfunction and heart failure with preserved ejection fraction (HFpEF) in particular (44,45), highlighting the potential parallel effects of inflammation on cardiac and pulmonary function. As a result of these shared pathophysiological pathways, we therefore hypothesized that accelerated decline in pulmonary function would predict incident cardiovascular events. Indeed, participants with rapid decline in lung function in our analysis demonstrated higher concentrations of hs-CRP—an established marker of systemic inflammation—compared with those without rapid decline. In the current study, FEV1 rapid decline in lung function was defined as the greatest quartile of decline, with FEV1 rapid decliners demonstrating an average decline of 120 ml/year, and FVC rapid decliners demonstrating an average 160 ml/year decline. These values are consistent with prior studies of rapid decline in pulmonary function (29). Both FEV1 rapid decline and FVC rapid decline were predictive of incident CVD independent of established cardiovascular risk factors and after accounting for the competing risk of noncardiovascular death.
Rapid decline in both FEV1 and FVC was associated with a higher risk of incident HF, even after adjusting for potentially confounding cardiovascular risk factors, including smoking status and accounting for pack-years, and NT-proBNP. In addition, the association persisted after excluding participants with incident CHD, showing that incident CHD or myocardial infarction does not explain this association. We are not aware of other studies demonstrating a relationship between changes in lung function and HF risk. Prior studies have shown a relationship between low baseline lung function and incident HF (45,46) and have suggested that higher levels of inflammatory markers identify persons with poor lung function who are at particularly high risk for HF (29). Although the HF phenotype in these studies (HFpEF vs. HF with reduced ejection fraction) is not clear, impaired lung function appears differentially predictive of HFpEF compared with HF with reduced ejection fraction (47), and the potential role for systemic inflammation in HFpEF has recently generated considerable interest (44,45). Activation of an inflammatory cascade may therefore be one mechanism linking lung function and HF, although further study is needed to support this hypothesis (46). Previous studies showed an association between inflammation and impaired pulmonary function (7). Indeed, in the current analysis, hs-CRP—a marker of systemic inflammation—was significantly higher among FEV1 and FVC rapid decliners than nonrapid decliners.
Rapid decline in FEV1 was associated with heightened risk of subsequent HF hospitalization predominantly during the early follow-up period, with an attenuated magnitude of association later in the follow-up period; this suggests that reverse causality may be present, such that FEV1 decline in some cases may be an early manifestation of HF. In contrast, rapid decline in FVC was associated with HF risk consistently throughout follow-up. One explanation for these findings is that rapid decline in FEV1 specifically may be largely an early manifestation of subclinical HF, with pulmonary congestion leading to interstitial and alveolar edema, consequent compression of the airways, and a rapid fall in FEV1 (16). Indeed, rapid decline in FEV1 secondary to early and undiagnosed HF (reserve causality) is expected to predict incident HF primarily during short-term follow-up, as observed in our analysis. In contrast, the consistent HF risk associated with rapid decline in FVC suggests a primary alteration in lung function predicting incident HF. Progressive deterioration in lung function may result in pulmonary hypertension and associated right HF when severe, although this is uncommon on a population level (48). An alternative, or additional, explanation is shared pathophysiological alterations including loss of elasticity in lung and vasculature (38), and possibly the myocardium, with resulting coupled impairments in cardiovascular and pulmonary reserve. For any degree of cardiovascular dysfunction, impaired pulmonary function may then predispose to manifestation of HF signs and symptoms. Future studies with serial echocardiography and spirometry will be necessary to interrogate this hypothesis.
We noted differential independent associations of rapid decline in FEV1 with the atherosclerotic endpoint of stroke. The association between decline in FEV1 and coronary disease mortality was previously demonstrated in the Copenhagen City Heart Study (14) and Baltimore Longitudinal Study of Aging (15). We also noted an association of rapid decline in FEV1 with incident CHD in unadjusted models and after adjustment for participant demographics. However, this association did not persist after further accounting for established cardiovascular risk factors. The reasons for these between-study differences are unclear but may relate to differences in the populations studied and in the adjustment variables between studies. To the extent that inflammation is 1 mechanism mediating the relationship between decline in pulmonary function and incident CVD, 1 potential explanation for this lack of association with CHD independent of established cardiovascular risk factors is that—for this endpoint at least—the additional contribution of pulmonary dysfunction to inflammation beyond established risk factors is modest. Similarly, in unadjusted analysis, both FEV1 rapid decline and FVC rapid decline were associated with higher risk of stroke. However, only rapid decline in FEV1 remained independently associated with incident stroke in fully adjusted models. Although reduced FEV1 has been demonstrated as related to incident stroke (49), to our knowledge, this is the first study demonstrating an association between rapid decline in FEV1 and stroke. Inflammation-associated pro-atherosclerotic mechanisms may help explain this association. In addition, decline in lung function is associated with hypertension (39), which is a major stroke risk factor and potential mediator (42). In our analysis, the association of rapid decline in FEV1 and incident stroke persisted after accounting for hypertension and blood pressure. Low pulmonary function has been associated with a higher risk of atrial fibrillation (50). Although the relationship between changes in lung function and incident atrial fibrillation is not known, a higher incidence of atrial fibrillation among the rapid decline group is also a possible explanation. Finally, rapid decline in lung function is associated with a higher incidence of COPD (13,36), which may in turn lead to higher levels of hematocrit and hemoglobin, and could predispose to stroke through elevation in the blood viscosity (51). However, we did not find any indication that hemoglobin is a mediator in the relationship between decline in FEV1 and stroke in our analysis (data not shown).
Current guidelines recommend serial spirometry for individuals at higher risk of developing pulmonary disease, such as heavy smokers and workers with an occupational exposure (20). Our results provide additional evidence that, in addition to providing information regarding risk of pulmonary disease, serial changes in spirometry also provides the clinician information regarding the risk of CVD. Our findings also demonstrate that rapid decline in FEV1 in particular predicts increased risk of incident HF. This risk was particularly high (4-fold increased) within 12 months, suggesting that clinicians should carefully consider incipient HF in patients with rapid changes in FEV1. Further studies are necessary to determine whether strategies to reduce the rate of decline in FEV1 and FVC reduce the incidence of CVDs.
The 3-year period between the 2 spirometry tests is relatively short, and test-retest variability may limit our ability to identify participants with actual accelerated decline in lung function impairment. However, spirometry is a reliable and reproducible test, with a coefficient of variation within subjects of about 5% (52), making variability between the 2 tests unlikely to bias our findings. Only pre-bronchodilator spirometric measurements are available in this study, although post-bronchodilator measures are also recommended by the Global Initiative for Chronic Obstructive Lung Disease (11). Finally, unmeasured and residual confounding of the relationship between change in lung function and cardiovascular outcomes could not be completely addressed by multivariate modeling.
Rapid decline in lung function over a 3-year period is associated with a heightened risk of incident CVD, including HF and stroke, independent of established cardiovascular risk factors and smoking. Rapid decline in FVC is associated with a heightened risk of HF throughout the approximately 17-year follow-up period, while rapid decline in FEV1 specifically predicts a 4-fold increase in the risk of incident HF at 1 year, and could be an early manifestation of HF.
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: In addition to predicting pulmonary disease, a rapid decline in spirometric measures of lung function, FEV1 and FVC is associated with a higher risk of CVD. Specifically, rapid decline in FEV1 is associated with a 4-fold increased risk of incident HF during the first 12 months of follow-up, suggesting that clinicians should be alert to manifestation of HF when patients exhibit rapid deterioration of FEV1.
TRANSLATIONAL OUTLOOK: Further studies are necessary to determine whether worsening of these indexes of pulmonary function reflect incipient HF and whether strategies to retard the decline in lung function can prevent or delay the onset of CVD.
The authors thank the staff and participants of the ARIC study for their important contributions.
The ARIC Study is carried out as a collaborative study supported by National Heart, Lung, and Blood Institute contracts (HHSN268201100005C, HHSN268201100006C, HHSN268201100007C, HHSN268201100008C, HHSN268201100009C, HHSN268201100010C, HHSN268201100011C, and HHSN268201100012C). The work for this paper was also supported by National Heart, Lung, and Blood Institute grants K08HL116792 and R01HL135008, American Heart Association grant 14CRP20380422, and the Brigham and Women’s Hospital Heart and Vascular Center Watkins Discovery Award (to Dr. Shah); Brazilian National Council for Scientific and Technological Development Grant 249481/2013-8 (to Dr. Nadruz); and the J.P. Lemann Foundation Jorge Paulo Lemann Harvard Medical School Cardiovascular Fellowship (to Dr. Silvestre). Dr. London is supported by the Intramural Research Program of the NIH, National Institute of Environmental Health Sciences (ZO1 ES043012). Dr. Shah has received consulting fees from Bellerophon Therapeutics; and has received research grants through the Brigham and Women's Hospital from Novartis. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- body mass index
- coronary heart disease
- chronic obstructive pulmonary disease
- cardiovascular disease
- forced expiratory volume in 1 s
- forced vital capacity
- heart failure
- Received May 29, 2018.
- Accepted June 18, 2018.
- 2018 American College of Cardiology Foundation
- GBD 2015 Mortality and Causes of Death Collaborators
- Roversi S.,
- Fabbri L.M.,
- Sin D.D.,
- Hawkins N.M.,
- Agusti A.
- Engström G.,
- Melander O.,
- Hedblad B.
- ↵World Health Organization. Chronic respiratory disease. Available at: http://www.who.int/gard/publications/chronic_respiratory_diseases.pdf. Accessed November 14, 2016.
- Gan W.Q.,
- Man S.F.P.,
- Senthilselvan A.,
- Sin D.D.
- Bonetti P.O.,
- Lerman L.O.,
- Lerman A.
- Janssens J.P.,
- Pache J.C.,
- Nicod L.P.
- ↵Global Initiative for Chronic Obstructive Lung Disease. GOLD 2017. Available at: https://goldcopd.org/gold-2017-global-strategy-diagnosis-management-prevention-copd/. Acessed July 23, 2018.
- National Heart, Lung, and Blood Institute of the National Institutes of Health
- Rosamond W.D.,
- Folsom A.R.,
- Chambless L.E.,
- et al.
- Folsom A.R.,
- Yamagishi K.,
- Hozawa A.,
- Chambless L.E.,
- for the Atherosclerosis Risk in Communities Study Investigators
- Howard G.,
- Wagenknecht L.E.,
- Cai J.,
- Cooper L.,
- Kraut M.A.,
- Toole J.F.
- Nadruz W.,
- Gonçalves A.,
- Claggett B.,
- et al.
- Michos E.D.,
- Selvin E.,
- Misialek J.R.,
- et al.
- Mannino D.M.,
- Davis K.J.
- Benjamin E.J.,
- Virani S.S.,
- Callaway C.W.,
- et al.
- Vestbo J.,
- Lange P.
- Hancox R.J.,
- Poulton R.,
- Greene J.M.,
- et al.
- Duprez D.A.,
- Hearst M.O.,
- Lutsey P.L.,
- et al.
- Vasan R.S.,
- Sullivan L.M.,
- Roubenoff R.,
- et al.
- Paulus W.J.,
- Tschöpe C.
- Shah S.J.,
- Kitzman D.W.,
- Borlaug B.A.,
- et al.
- Lam C.S.P.,
- Lyass A.,
- Kraigher-Krainer E.,
- et al.
- Seeger W.,
- Adir Y.,
- Barbera J.A.,
- et al.
- Li J.,
- Agarwal S.K.,
- Alonso A.,
- et al.
- Laszlo G.