Association of Chronic Respiratory Symptoms With Incident Cardiovascular Disease and All-Cause Mortality

Abstract

Background

Respiratory and cardiovascular diseases (CVDs) frequently coexist; however, there is limited evidence on the relationship between chronic respiratory symptoms in young adulthood and late-onset CVD.

Research Question

Are chronic respiratory symptoms in young adulthood associated with CVD and all-cause mortality in later life?

Study Design and Methods

A total of 4,621 participants from the Coronary Artery Risk Development in Young Adults Study (CARDIA) cohort study aged 18 to 30 years were included. Chronic respiratory symptoms were identified through respiratory symptom questionnaires in two consecutive examinations. Incident CVD and all-cause mortality were adjudicated over 30-year follow-up. Multivariable Cox proportional hazards models were used to explore the association of chronic respiratory symptoms with incident CVD and all-cause mortality.

Results

During a median follow-up of 30.9 years, 284 CVD events (6.15%) and 378 deaths (8.18%) occurred. Following multivariable adjustment for demographic characteristics, cardiovascular risk factors, smoking, and lung function, the hazard ratios (95% CIs) for CVD events were 1.51 (1.18-1.93) for any respiratory symptom, 1.57 (1.18-2.09) for cough or phlegm, 1.31 (1.01-1.68) for wheeze, 1.73 (1.25-2.41) for shortness of breath, and 1.32 (1.01-1.71) for chest illnesses. Similar findings were also observed in all-cause mortality. Comparing zero vs three to four respiratory symptoms, the hazard ratios (95% CIs) were 1.97 (1.34-2.91) for CVD and 1.75 (1.23-2.47) for all-cause mortality. Similar results were observed in various sensitivity analyses.

Interpretation

Chronic respiratory symptoms in young adulthood are associated with an increased risk of CVD and all-cause mortality in midlife independent of established cardiovascular risk factors, smoking, and lung function. Identifying chronic respiratory symptoms in young adulthood may help provide prognostic information regarding future cardiovascular health.

Clinical Trial Registration

ClinicalTrials.gov; No.: NCT00005130; URL: https://www.clinicaltrials.gov

Graphical Abstract

FOR EDITORIAL COMMENT, SEE PAGE 876

Take-home Points.

Study Question: Are chronic respiratory symptoms in young adulthood associated with CVD and all-cause mortality in later life?

Results: Chronic respiratory symptoms in young adulthood are associated with an increased risk of CVD and all-cause mortality in midlife independent of established cardiovascular risk factors, smoking, and lung function.

Interpretation: Identifying chronic respiratory symptoms in young adulthood may help provide prognostic information regarding future cardiovascular health.

Respiratory and cardiovascular diseases (CVDs) frequently coexist and share common risk factors.^{1, 2, 3} Respiratory symptoms, the prodromal stage of respiratory disease, can be used to identify individuals earlier in the disease process who may have an increased risk of reduced lung function and pulmonary disease later in life.⁴ Although 21% of adults reported at least one respiratory symptom,⁵ little is known regarding the clinical relevance of respiratory symptoms in cardiovascular outcomes among young adults.

Previous studies suggested that chronic respiratory symptoms may cause activation of systemic inflammation,^6,7 endothelial dysfunction,⁸ and impaired vascular reactivity.⁹ These changes are common antecedents of CVD. Epidemiologic studies reported that respiratory symptoms were associated with CVD events. These studies were conducted mainly in middle-aged and older adults who usually experience confounding by symptomatic disease or treatment-related influences.^10,11 Little evidence currently exists regarding the association between chronic respiratory symptoms in young adulthood and cardiovascular outcomes in midlife. Furthermore, identifying the contribution of chronic respiratory symptoms early in life, prior to the onset of CVD and related risk factors, may inform our understanding of how dysfunction in respiratory system affects long-term CVD progression and prognosis.

The Coronary Artery Risk Development in Young Adults (CARDIA) study participants were initially young adults aged 18 to 30 years, which makes it well suited for investigating the relationship between early respiratory symptoms in generally healthy young adults and cardiovascular outcomes over the next 30 years. Kalhan et al⁴ used the same database and the same respiratory symptoms, with a different outcome, namely lung function and lung disease. Another study, from Cuttica et al,¹² showed the association of lung function with subsequent cardiovascular events. These studies together lead to a strong expectation that respiratory symptoms are related to future disease events in the CARDIA study; however, whether these associations are independent from lung function remains unclear. We hypothesized that the presence of chronic respiratory symptoms in young adulthood is associated with an increased risk of incident CVD and all-cause mortality in midlife independent of CVD risk factors, smoking, and lung function.

Study Design and Methods

Study Population

CARDIA is a multicenter longitudinal cohort study of 5,115 healthy young adults aged 18 to 30 years recruited in 1985 to 1986 from four US metropolitan communities (Birmingham, Alabama; Chicago, Illinois; Minneapolis, Minnesota; and Oakland, California). A detailed design of CARDIA has been published previously.¹³ Participants have been contacted via telephone annually and invited to participate in serial follow-up examinations at years 2, 5, 7, 10, 15, 20, 25, and 30 following baseline. The retention rate was high, with 71% of surviving participants completing the year 30 (2015-2016) examination. All participants provided written informed consent at each examination, and the institutional review board at each study site and coordinating center approved the study procedure for all examinations.

From the CARDIA cohort enrolled at baseline, participants were included who completed respiratory symptom questionnaires at baseline and year 2 examinations and have assessment of end point events over the 30-year follow-up. We excluded participants who withdrew study consent (n = 1), had a myocardial infarction at baseline examination (n = 1), and did not have information for respiratory symptom record at baseline or year 2 examination (n = 492). The final sample size for the analysis of chronic respiratory symptoms and cardiovascular outcomes was 4,621 participants (e-Fig 1).

Assessment of Respiratory Symptoms

Respiratory symptoms, including cough or phlegm, wheezing, shortness of breath, and chest illnesses, were determined by using questions adapted from the American Thoracic Society respiratory questionnaire.¹⁴ Detailed definitions of each respiratory symptom are documented in e-Table 1. As previously described,⁴ any symptom was defined by any one symptom category being reported at both the baseline and year 2 examinations (the same symptom being reported at both examinations). The participants were classified further according to the number of chronic respiratory symptoms using a score ranging from 0 to 4. In this classification, a score of 0 indicated no chronic respiratory symptom, and the scores 1 to 4 indicated the number of chronic respiratory symptoms.

Ascertainment of Covariates

Standardized questionnaires and protocols were used to collect data on participant demographic characteristics, smoking status (defined as never, former, and current), alcohol consumption, and physical activity.¹³ Height and weight were measured without shoes and in light clothing and were used to calculate BMI in kilogram per meter squared. BP was measured in triplicate following a 5-min rest; the average of the second and third measurements was used. FEV₁ and FVC were measured with at least five satisfactory maneuvers. The maximum of prebronchodilator FEV₁ and FVC were used. Standard spirometry procedures as recommended by the American Thoracic Society were followed at all of the examinations.^15,16 Other laboratory measures, such as fasting glucose, total cholesterol, low-density lipoprotein cholesterol (LDL-C), and high-density lipoprotein cholesterol (HDL-C), were collected for analysis.

For time-varying categorical covariates such as smoking, we calculated cumulative measures from 1985 to 2015. Participants were coded as always, sometimes, or never having the condition or behavior from 1985 to 2015. For continuous covariates, which could vary over time (including BMI, alcohol consumption, physical activity, BP, fasting glucose, LDL-C, HDL-C, and lung function), we calculated the average for each of these measures from 1985 to 2015.¹⁷

Cardiovascular Outcomes

Participants were contacted annually to identify the possible outcomes through December 2016. The primary outcome was fatal and nonfatal CVD events, including coronary heart disease, cerebrovascular disease, and other heart or vascular diseases, specifically fatal and nonfatal myocardial infarction, acute coronary syndrome, hospitalization for heart failure, intervention for peripheral arterial disease, stroke, transient ischemic attack, or death from cardiovascular causes.^17,18 All-cause mortality (all cases of death) was examined as a secondary outcome. For each outcome, medical records, death certificates, informant interviews (for outpatient deaths), and autopsy reports, when available, were adjudicated independently by two reviewers of the Endpoint Committee. Definitions of each outcome have been previously described.^{19, 20, 21}

Statistical Analysis

Baseline characteristics are expressed as the mean ± SD or number (%). Participants were divided into four groups according to the number of chronic respiratory symptoms. Missing covariate values were generated by using multiple imputation with 10 iterations.^22,23 The percentage of missing data before imputation was 5% for lung function (FEV₁, FVC, and FEV₁/FVC ratio), 1.4% for fasting glucose, and < 1% for other covariates. Censored subjects were those who had not experienced CVD or all-cause mortality during follow-up. The incidence rate of CVD and all-cause mortality was calculated by dividing the number of incident cases by the total follow-up duration (person-years). The survival probability of primary outcomes according to the presence of respiratory symptoms was shown by using unadjusted Kaplan-Meier curves.

Cox proportional hazards assumption tests were performed for each model by using the Schoenfeld residuals test. There was no significant departure from proportionality in hazards over time. The hazard ratio (HR) and 95% CI for CVD events and all-cause mortality were analyzed by using the multivariable Cox proportional hazards regression models. Model 1 was unadjusted. Model 2 adjusted for baseline age, sex, race, BMI, smoking status, alcohol consumption, physical activity, systolic BP, fasting glucose, LDL-C, and HDL-C. Model 3 adjusted for baseline age, sex, race, and cumulative risk factors (BMI, smoking status, alcohol consumption, physical activity, systolic BP, fasting glucose, LDL-C, and HDL-C). Model 4 adjusted for variables in Model 3 plus cumulative lung function.

Statistical significance for a trend of the HRs across the number of chronic respiratory symptoms was tested. The potential effect modification by sex, race, BMI, and smoking status was estimated by using stratified analysis (with an accompanying test for statistical interaction). A sensitivity analysis was performed to exclude participants with known respiratory disease, including chronic bronchitis, emphysema, asthma, and COPD, at baseline. To account for the possible influence of incident respiratory disease during the follow-up period, a sensitivity analysis was performed to exclude these participants with incident respiratory disease. A two-sided P value < .05 was considered statistically significant. All analyses were conducted by using SPSS version 20 (IBM SPSS Statistics, IBM Corporation).

Results

Participant Characteristics

Baseline characteristics of the 4,621 participants are presented in Table 1. Participants with more chronic respiratory symptoms were more likely to be female and less physically active. Overall, 65%, 55%, and 36% of never, former, and current smokers had zero symptoms, whereas 9%, 13%, and 21% had two symptoms, and 3%, 4%, and 14% had three to four symptoms. The highest baseline alcohol consumption, BMI, and glucose values, as well as the lowest HDL-C levels, were observed in participants with three or four chronic respiratory symptoms. Lung function, such as FEV₁, FVC, and FEV₁/ FVC ratio, gradually decreased with the number of chronic respiratory symptoms but was still in the normal range. The difference in lung function expanded greatly over the years in those with symptoms early in the study.⁴ However, age, race, BP, total cholesterol, and LDL-C levels were not significantly different as stratified according to the number of chronic respiratory symptoms.

Table 1.

Baseline Characteristics of Participants Stratified According to Number of Chronic Respiratory Symptoms

Variable	No. of Chronic Respiratory Symptoms				P Value
	0	1	2	3 or 4
No.	2,525	1,182	611	303
Age, y	24.8 ± 3.6	25.1 ± 3.6	24.9 ± 3.5	25.3 ± 3.5	.10
Female, No. (%)	1,335 (52.9)	653 (55.2)	353 (57.8)	191 (63.0)	.002
Black, No. (%)	1,268 (50.2)	560 (47.4)	296 (48.4)	162 (53.5)	.19
Smoking status, No. (%)
Never	1,690 (66.9)	608 (51.4)	240 (39.3)	92 (30.4)	< .001
Former	343 (13.6)	179 (15.1)	78 (12.8)	24 (7.9)	.01
Current	492 (19.5)	395 (33.4)	293 (48.0)	187 (61.7)	< .001
Alcohol consumption, mL/d	10.3 ± 19.7	12.3 ± 21.4	16.0 ± 27.0	16.7 ± 24.5	< .001
Height, cm	170.6 ± 9.3	170.5 ± 9.2	169.7 ± 9.3	169.0 ± 9.1	.006
BMI, kg/m2	24.1 ± 4.5	24.5 ± 4.8	24.7 ± 5.1	26.2 ± 6.2	< .001
SBP, mm Hg	110.4 ± 11.0	110.5 ± 10.7	109.8 ± 11.0	110.0 ± 11.1	.46
DBP, mm Hg	68.7 ± 9.5	68.6 ± 9.3	68.1 ± 10.0	68.8 ± 10.2	.54
Physical activity, exercise units	426.6 ± 302.9	418.2 ± 296.8	420.1 ± 299.1	372.6 ± 277.8	.03
FEV1, mL	3,601.4 ± 770.9	3,542.4 ± 815.5	3,403.1 ± 781.4	3,278.0 ± 779.4	< .001
FVC, mL	4,319.2 ± 1,005.2	4,305.3 ± 1,065.1	4,187.4 ± 1,006.4	4,075.3 ± 1,016.9	< .001
FEV1/FVC ratio, %	83.9 ± 6.0	82.9 ± 6.6	81.8 ± 7.2	81.0 ± 7.7	< .001
Asthma history, n (%)	74 (2.9)	139 (11.8)	117 (19.1)	62 (20.5)	< .001
Glucose, mg/dL	82.2 ± 12.0	82.5 ± 15.8	82.0 ± 8.9	87.8 ± 38.7	< .001
Total cholesterol, mg/dL	176.7 ± 32.7	176.9 ± 34.0	175.7 ± 34.5	180.8 ± 34.2	.17
HDL-C, mg/dL	53.6 ± 12.8	52.6 ± 13.0	52.3 ± 13.5	51.7 ± 14.8	.008
LDL-C, mg/dL	109.0 ± 30.7	109.4 ± 31.7	108.6 ± 32.0	112.0 ± 31.9	.44

Data are presented as mean ± SD unless otherwise indicated. Normal ranges as follows: fasting glucose, 70-110 mg/dL; total cholesterol, < 200 mg/dL; HDL-C, > 40 mg/dL; and LDL-C, < 130 mg/dL. DBP = diastolic BP; HDL-C = high-density lipoprotein cholesterol; LDL-C = low-density lipoprotein cholesterol; SBP = systolic BP.

CVD and Chronic Respiratory Symptoms

During a median 30.9-year follow-up, 378 deaths (8.18%) and 284 CVD events (6.15%) occurred in the entire cohort. The number of chronic respiratory symptoms was linearly related to the outcome measures (Table 2). The incidence rates of both CVD (from 1.53 to 4.32 per 1,000 person-years, P value for trend < .001) and all-cause mortality (from 2.17 to 4.86 per 1,000 person-years, P value for trend < .001) were gradually increased with an increasing number of chronic respiratory symptoms. The unadjusted and adjusted effect estimates for the association of chronic respiratory symptoms with CVD and all-cause mortality are shown in Table 2. Compared with the group without chronic respiratory symptoms (reference group), the group with three or four chronic respiratory symptoms had a significantly higher risk for CVD (HR, 2.89; 95% CI, 2.00-4.16) and all-cause mortality (HR, 2.26; 95% CI, 1.62-3.16). Following adjustment for baseline demographic characteristics and CVD risk factors, HRs for CVD and all-cause mortality increased continuously with an increasing number of respiratory symptoms. These associations remained significant following adjustment for cumulative CVD risk factors. In the final, fully adjusted Cox model, for the highest category in chronic respiratory symptom compared with the lowest category, the hazard of CVD increased by 97% (HR, 1.97; 95% CI, 1.34-2.91) and all-cause mortality by 75% (HR, 1.75; 95% CI, 1.23-2.47) following further adjustment for cumulative lung function. Taken together, a depiction of chronic respiratory symptoms with CVD and all-cause mortality is presented in e-Figure 2.

Table 2.

Hazard Ratios and 95% CIs of Cardiovascular Disease and All-Cause Mortality According to Number of Chronic Respiratory Symptoms

Outcomes	Events (No.)	Follow-Up Duration (Person-Year)	Incidence Rate (Per 1000 Person-Year)	Hazard Ratios (95% CI)
				Model 1	Model 2	Model 3	Model 4
Cardiovascular disease
0	116	75,689	1.53	1 (ref.)	1 (ref.)	1 (ref.)	1 (ref.)
1	81	34,833	2.33	1.53 (1.15-2.03)	1.44 (1.07-1.93)	1.35 (1.01-1.80)	1.33 (1.00-1.78)
2	49	17,864	2.74	1.81 (1.30-2.53)	1.73 (1.22-2.46)	1.67 (1.19-2.36)	1.64 (1.16-2.31)
3 or 4	38	8,804	4.32	2.89 (2.00-4.16)	2.15 (1.44-3.19)	2.02 (1.37-2.97)	1.97 (1.34-2.91)
P for trend				< .001	< .001	.001	.002
All-cause mortality
0	166	76,426	2.17	1 (ref.)	1 (ref.)	1 (ref.)	1 (ref.)
1	96	35,401	2.71	1.26 (1.00-1.61)	1.17 (0.90-1.51)	1.14 (0.88-1.47)	1.12 (0.86-1.44)
2	72	18,146	3.97	1.83 (1.39-2.42)	1.60 (1.20-2.13)	1.67 (1.26-2.21)	1.64 (1.23-2.18)
3 or 4	44	9,052	4.86	2.26 (1.62-3.16)	1.61 (1.13-2.30)	1.77 (1.25-2.51)	1.75 (1.23-2.47)
P for trend				< .001	.004	< .001	.001

Model 1 = unadjusted. Model 2 = adjusted for baseline age, sex, race, BMI, smoking status, alcohol consumption, physical activity, systolic BP, fasting glucose, low-density lipoprotein cholesterol, and high-density lipoprotein cholesterol. Model 3 = adjusted for baseline age, sex, race, and cumulative risk factors (BMI, smoking status, alcohol consumption, physical activity, systolic BP, fasting glucose, low-density lipoprotein cholesterol, and high-density lipoprotein cholesterol). Model 4 = adjusted for variables in model 3 plus cumulative lung function (FEV₁, FVC, or FEV₁/ FVC). Ref. = reference.

The frequencies of reporting each respiratory symptom at baseline and year 2 examinations are presented (e-Table 2). For each symptom, an incrementally higher risk of CVD and all-cause mortality was observed for the presence of symptoms compared with the absence of symptoms (e-Fig 3, Fig 1). The associations between the presence of each symptom and outcomes were confirmed after adjusting for demographic characteristics, cumulative CVD risk factors, and lung function. For the presence of any symptom compared with those without symptoms, the risk of CVD increased by 51% (HR, 1.51; 95% CI, 1.18-1.93) (Fig 2A) and all-cause mortality by 36% (HR, 1.36; 95% CI, 1.10-1.68) (e-Fig 4A). For the presence of cough or phlegm compared with those without the respective symptom, the risk of CVD increased by 57% (HR, 1.57; 95% CI, 1.18-2.09) (Fig 2B) and all-cause mortality by 35% (HR, 1.35; 95% CI, 1.05-1.75) (e-Fig 4B). For the presence of wheeze compared with the absence of this symptom, the risk of CVD increased by 31% (HR, 1.31; 95% CI, 1.01-1.68) (Fig 2C) and all-cause mortality by 44% (HR, 1.44; 95% CI, 1.16-1.78) (e-Fig 4C). For the presence of shortness of breath compared with the absence of this symptom, the risk of CVD increased by 73% (HR, 1.73; 95% CI, 1.25-2.41) (Fig 2D) and all-cause mortality by 57% (HR, 1.57; 95% CI, 1.16-2.12) (e-Fig 4D). For those with chest illnesses compared with those without respective symptoms, the risk of CVD increased by 32% (HR, 1.32; 95% CI, 1.01-1.71) (Fig 2E) and all-cause mortality by 32% (HR, 1.32; 95% CI, 1.06-1.66) (e-Fig 4E).

Figure 1.

Kaplan-Meier curves for each respiratory symptom and CVD. Kaplan-Meier curves are used to illustrate the association of any symptom (A), cough or phlegm (B), wheeze (C), shortness of breath (D), and chest illnesses (E) with CVD over 30 years’ follow-up. P values were calculated by using the log-rank test. CVD = cardiovascular disease.

Figure 2.

Association of the presence of chronic respiratory symptom with cardiovascular disease. Model 1: unadjusted. Model 2: adjusted for baseline age, sex, race, BMI, smoking status, alcohol consumption, physical activity, systolic BP, fasting glucose, low-density lipoprotein cholesterol, and high-density lipoprotein cholesterol. Model 3: adjusted for baseline age, sex, race, and cumulative BMI, smoking status, alcohol consumption, physical activity, systolic BP, fasting glucose, low-density lipoprotein cholesterol, and high-density lipoprotein cholesterol. Model 4: adjusted for variables in model 3 plus cumulative lung function (FEV₁, FVC, or FEV₁/FVC). HR = hazard ratio.

Subgroup Analysis

We performed a stratified analysis according to sex, race, and BMI categories (e-Fig 5). The risks of all-cause mortality and CVD increased with an increasing number of chronic respiratory symptoms in all study subgroups. The associations between the number of chronic respiratory symptoms and CVD and all-cause mortality were similar in various subgroups (all P for interaction > .1). To account for a potential influence of smoking, we performed another subgroup analysis stratified according to smoking status (Fig 3). A higher HR for CVD (P for interaction < .001) but not for all-cause mortality (P for interaction > .1) was observed in the smoking subgroup.

Figure 3.

Subgroup analyses of association between the number of chronic respiratory symptoms and cardiovascular disease and all-cause mortality stratified according to smoking status. Nonsmoking was determined as never or former smokers; smoking was determined as current smoker at year 0. HRs were adjusted for age, sex, race, and cumulative BMI, smoking status, alcohol consumption, physical activity, systolic BP, fasting glucose, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, and lung function. HR = hazard ratio.

Sensitivity Analysis

Excluding participants with known respiratory disease at baseline, including chronic bronchitis, emphysema, asthma, and COPD, produced nearly identical results. These results showed that the number of chronic respiratory symptoms was an independent predictor of CVD and all-cause mortality even in participants without known respiratory disease at baseline (e-Table 3). During follow-up, 708 subjects (15.3% of the study population) developed chronic bronchitis, emphysema, asthma, and COPD. Analyses censoring these subjects with incident respiratory disease (e-Table 4) or censoring those without lung function assessment (e-Table 5) also produced similar results. In addition, defining chronic respiratory symptoms over the 10-year period (ie, symptom consistently reported at year 0, year 2, and year 10 examinations) yield similar results (e-Table 6).

Discussion

In this population-based cohort study of generally healthy young adults with a median 30.9-year follow-up, we found that chronic respiratory symptoms in young adulthood, including cough or phlegm, wheezing, shortness of breath, and chest illnesses, were associated with a higher risk for CVD and all-cause mortality in midlife. We also found a graded association between the number of chronic respiratory symptoms and the primary outcomes. These associations remained significantly associated following multivariable adjustment for demographic characteristics, cumulative CVD risk factors, and lung function. The findings suggest that the presence of chronic respiratory symptoms may serve as a prognostic surrogate marker for future cardiovascular event and mortality.

Prior studies of respiratory symptoms and primary outcomes have reported conflicting results. No association between respiratory symptoms and cardiovascular outcomes was reported in the Nord-Trøndelag Health Study (HUNT) study.¹¹ In contrast, findings from an urban community study showed significant association between respiratory symptoms and cardiovascular outcomes.²⁴ In the Australian Longitudinal Study of Ageing (ALSA), older participants who reported significant respiratory symptoms had a greater risk of mortality than asymptomatic subjects.²⁵ These studies have included older individuals, among whom subclinical and overt cardiovascular and metabolic abnormalities may already be established and may contribute to the outcomes. Considering that, young adulthood is a relatively early critical period to investigate the nature effect of respiratory symptoms on CVD development. However, few studies have specifically explored this relationship in relatively young individuals. In the present study, we show that the presence of chronic respiratory symptoms in young adulthood increases future risk of CVD and all-cause mortality.

Although respiratory symptoms are commonly reported in individuals across all fields of clinical care, no valid criteria are available to date to distinguish between transient and chronic symptoms at an early stage. Moreover, it is unclear if how long the symptoms last has prognostic value for the adverse outcomes later in life. The required durability of symptoms in respiratory disease such as chronic bronchitis is at least 2 consecutive years.²⁶ The CARDIA study found that the symptoms reported at both baseline and 2 years later reflect a true biologic process and not transient phenomena in the development of lung disease.⁴ Consistently, our findings also showed that chronic symptoms in two consecutive years (ie, presence at both baseline and year 2) has greater prognostic significance in CVD and all-cause mortality than simply having these symptoms at a single point in time (e-Table 7). Considering that, in this study, the respiratory symptoms consistently reported at both baseline and year 2 examinations (ie, presence in two consecutive years) were considered chronic.

It is well-known that current smoking and physician-diagnosed asthma are among the most common causes of respiratory symptoms in relatively young cohorts. In this study, however, a considerable proportion of participants reporting symptoms did not report smoking or asthma. Moreover, the association between respiratory symptoms and primary outcomes persisted following adjustment for smoking status or censoring subjects with asthma. It indicated that other potential factors might influence the presence of respiratory symptoms, including premature birth, perinatal exposure to smoking, childhood respiratory infection, and air pollution exposure.^{27, 28, 29, 30}

We raise an interesting question regarding the significance of chronic respiratory symptoms in the development of CVD. What, then, could be the possible mechanisms for why chronic respiratory symptoms in young adulthood are associated with long-term CVD outcomes? Impaired lung function may be an important mediator linking chronic symptoms and future CVD risk. A study from the same database as our report showed that chronic respiratory symptoms in young adulthood, even in individuals with normal spirometry results, accelerated decline in lung function.⁴ Reduced lung function significantly increased the incidence of subsequent CVD and mortality.^3,31 In addition, persistent activation of systemic inflammation may be another mechanism implicated in this process. Inflammatory markers such as C-reactive protein and various chemokines are increased in individuals with chronic respiratory symptoms.^6,7 Increased levels of these inflammatory markers have a significant effect on the development of atherosclerosis, myocardial infarction, coronary heart disease, and other cardiovascular events.³² Lastly, chronic symptoms are associated with systemic endothelial dysfunction and impaired vascular reactivity, which also promote subsequent CVD and mortality.^8,9

The current study has several strengths. First, this study was, to the best of our knowledge, the first to investigate the association of chronic respiratory symptoms with CVD and all-cause mortality in a relatively large cohort of generally healthy young adults using a well-established and validated database over 30-year follow-up. Second, because the CARDIA study focuses specifically on CVD risks, the standardized outcome adjudication protocol was designed to ensure the accuracy of outcome ascertainment.¹⁷ Third, we performed a series of sensitivity analyses. Consistency between the results of the primary analyses and the sensitivity analyses provides convincing evidence and further increases the reliability of our findings. However, some potential limitations should be considered. First, participants were not included in these analyses due to missing symptom assessment at the baseline or year 2 examination. These participants who were excluded may have affected the study findings, although the retention of CARDIA participants remained high. Second, the reporting symptoms may be transient phenomena and probably change over time in individuals. To address this factor, the same symptom consistently reported at both the baseline and year 2 examinations was considered chronic. Third, with the development of incident respiratory disease, the association between respiratory symptoms and outcomes may be influenced. To minimize this possibility, we performed a sensitivity analysis to exclude those with incident respiratory disease occurring during follow-up, and this analysis also showed similar results. Fourth, data on the detailed cause of death may be likely to provide further insight into how these symptoms affect mortality. Finally, this was an observational study, and other potential residual and unknown confounding factors may be present.

Interpretation

Chronic respiratory symptoms in young adulthood were associated with an increased risk of CVD and all-cause mortality in midlife independent of established cardiovascular risk factors, smoking, and lung function. It suggests that identifying chronic respiratory symptoms in young adulthood may help provide prognostic information regarding future cardiovascular health.