Lung Function and Respiratory Symptoms in Association with Mortality: The HUNT Study

Abstract

Whether respiratory symptoms are associated with mortality independent of lung function is unclear. The authors explored the association of the exposures i) lung function, ii) respiratory symptoms, and iii) lung function and respiratory symptoms combined, with the outcomes all-cause and cardiovascular mortality. The study included 10,491 adults who participated in the Nord-Trøndelag Health Study (HUNT) Lung Study in 1995–1997 and were followed through 2009. Cox regression was used to calculate adjusted hazard ratios (HRs) with 95% confidence intervals for all-cause and cardiovascular mortality associated with pre-bronchodilator% predicted forced expiratory volume in 1 second (ppFEV₁), chronic obstructive pulmonary disease (COPD) grades, and respiratory symptoms (chronic bronchitis, wheeze, and levels of dyspnoea). Lung function was inversely associated with all-cause mortality. Compared to ppFEV₁ ≥100, ppFEV₁ <50 increased the HR to 6.85 (4.46–10.52) in women and 3.88 (2.60–5.79) in men. Correspondingly, compared to normal airflow, COPD grade 3 or 4 increased the HR to 6.50 (4.33–9.75) in women and 3.57 (2.60–4.91) in men. Of the respiratory symptoms, only dyspnoea when walking remained associated with all-cause mortality after controlling for lung function (HR 1.73 [1.04–2.89] in women and 1.57 [1.04–2.36] in men). Analyses of lung function and dyspnoea when walking as a combined exposure further supported this finding. Overall, associations between lung function and cardiovascular mortality were weaker, and respiratory symptoms were not associated with cardiovascular mortality. In conclusion, lung function was inversely associated with all-cause and cardiovascular mortality, and dyspnoea when walking was associated with all-cause mortality independent of lung function.

Keywords :

• Cardiovascular
• Cough
• Dyspnoea
• Epidemiology
• Forced expiratory volume
• Wheeze

Introduction

Chronic obstructive pulmonary disease (COPD) is primarily characterised by persistent airflow limitation (1). However, undiagnosed people with COPD may not seek medical advice until they experience respiratory symptoms like cough, wheeze, or dyspnoea, being unaware of possible lung function impairment (1, 2). COPD may reach an advanced stage before such symptoms become prominent, and not all patients experience respiratory symptoms (1). In addition, limited awareness of COPD among medical doctors and underuse of spirometry may contribute to the underdiagnosis of COPD (1–3). Both lung function level and presence of respiratory symptoms are assumed to be important for the prognosis of patients with COPD (1). Although the association between lung function and mortality has been thoroughly studied, the association between respiratory symptoms and mortality is less clear.

During the last three decades, several epidemiological studies have found associations between impaired lung function and increased all-cause or cardiovascular (CV) mortality (4–19). Respiratory symptoms have been found to be associated with all-cause and CV mortality in some studies where lung function has not been accounted for (20–23). However, whether respiratory symptoms are associated with all-cause and CV mortality independent of lung function needs to be further explored as previous literature is inconclusive (11–19, 24).

In this large population-based study we therefore aimed to explore the association of the exposures i) lung function, ii) respiratory symptoms, and iii) lung function and respiratory symptoms combined, with the outcomes all-cause and CV mortality.

Methods

Study population

All residents in Nord-Trøndelag County in Norway aged 19 years or older were invited to participate in the second survey of the Nord-Trøndelag Health Study (HUNT2) from August 1995 to June 1997. HUNT2 has been described elsewhere (25–27). Among the 65,237 participants in HUNT2 (69.5% of those invited), a 5% random sample and a symptom sample were invited to participate in the Lung Study (28, 29). Briefly, the symptom sample included participants reporting attacks of wheezing or breathlessness during the last 12 months, having ever had asthma, and/or having ever used asthma medication, and who were not included in the random sample.

Among the 10,693 participants eligible for the current study, we excluded 202 participants with missing information on body mass index (BMI), systolic blood pressure (SBP), total cholesterol, diabetes mellitus (DM), and cardiovascular disease (CVD). For the analyses of CV mortality we additionally excluded 1029 participants with CVD (angina pectoris, myocardial infarction, or stroke) at baseline. The remaining 10,491 participants for the all-cause mortality analyses and 9462 participants for the CV mortality analyses were weighted to represent 51,854 and 48,355 people, respectively, from the general population of Nord-Trøndelag County. This weighting has been described in detail elsewhere (28). A flow chart of inclusion and exclusion in the Lung Study is presented in Figure 1.

Figure 1.  Flow chart of inclusion and exclusion in the Lung Study.

Study variables

All participants in HUNT2 provided information about life style factors, complaints, diseases, and demographics (25, 26). Participants in the Lung Study completed an additional lung specific questionnaire and an interview, and performed flow volume spirometry (29).

Lung function

Flow volume spirometry was recorded according to the 1994 American Thoracic Society recommendations (30) as described elsewhere (29). Pre-bronchodilator forced vital capacity (FVC) and forced expiratory volume in one second (FEV₁) were obtained, and local prediction equations were used (29). Lung function was classified in two ways;% predicted FEV₁ (ppFEV₁) was categorised as ≥100, 80–99, 50–79, and <50; pre-bronchodilator COPD was defined as FEV₁/FVC <70% and airflow limitation was graded according to modified Global Initiative for Chronic Obstructive Lung Disease (GOLD) criteria as grade 1 (ppFEV₁ ≥80), grade 2 (50≤ ppFEV₁ <80), and grade 3 or 4 (ppFEV₁ <50) (1). Normal was defined as FEV₁/FVC ≥70% and% predicted FVC ≥80. Participants with possible restrictive lung function impairment (FEV₁/FVC ≥70% and% predicted FVC <80) were excluded from the COPD analyses as they would otherwise have been included in the “normal” category.

Respiratory symptoms

Chronic bronchitis (4.6% missing data) was measured by a question included in the main questionnaire: “Have you had a cough with phlegm for periods of at least three months during each of the last two years?” Wheeze (0.4% missing data) was measured by a question included in the interview: “During the last 12 months, have you had wheezing in your chest at any time?” Possible answers were “yes” or “no”. The lung specific questionnaire included four questions about dyspnoea during various activities from the Norwegian Respiratory Questionnaire (31), and these questions were used to generate a dyspnoea scale (25.3% missing data) (28), which is presented in eTable 1 in the web appendix. This scale included dyspnoea when walking uphill, climbing stairs, walking flat, and sitting still. The dyspnoea scale was dichotomized at dyspnoea when walking when counting number of respiratory symptoms. Additionally, we created combined exposures of lung function and each of chronic bronchitis, wheeze, and dyspnoea when walking or sitting. Participants with missing data on respiratory symptoms were excluded from the analyses including respiratory symptoms. To assess potential selection bias due to missing data we compared baseline characteristics and associations with all-cause and CV mortality between participants with and without data on the dyspnoea scale.

Table 1.  Statistical models for the association of lung function and respiratory symptoms with all-cause and cardiovascular mortality

Exposure	Model 1	Model 2	Model 3	Model 4
Lung function	Age	Model 1 + smoking, education		Model 2 + BMI, physical activity, CVD, DM, SBP, total cholesterol
Respiratory symptoms	Age	Model 1 + smoking, education, BMI, physical activity, CVD	Model 2 + lung function	Model 3 + DM, SBP, total cholesterol
Lung function combined with respiratory symptoms	Age	Model 1 + smoking, education, BMI, physical activity, CVD		Model 2 + DM, SBP, total cholesterol

Abbreviations: BMI = body mass index, CVD = cardiovascular disease, DM = diabetes mellitus, SBP = systolic blood pressure.

Follow-up

The participants were followed from the date of attendance in HUNT2 to the date of death or until the end of follow-up, December 31 2009, whichever came first. The primary endpoints were all-cause and CV death (International Classification of Diseases 10: I00–I99). Date and cause of death were obtained from the Cause of Death Registry in Norway which is practically complete due to the unique personal identification number of all residents in Norway.

Statistical analyses

To enhance the generalizability of the results the regression analyses were weighted with an inverse probability weight (32) to reflect the distribution in the general population (28). Sex specific (33) statistical analyses were performed with few exceptions.

Deaths per 1000 person-years of follow-up were calculated for categories of ppFEV₁, COPD grades, and respiratory symptoms. Cox proportional hazard regression was used to compute adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) for the association of lung function and respiratory symptoms with all-cause and CV mortality.

Possible confounders included in the Cox proportional hazard models were age (<40, 40–49, …, ≥80 years), smoking (current, former, never, unknown [5.5%]), education (<10, 10–12, ≥13 years, unknown [4.9%]), BMI (<18.5, 18.5–24.9, 25.0–29.9, ≥30.0 kg/m²), physical activity (inactive, light activity <1 hours/week, light activity 1–2 hours/week, light activity ≥3 hours/week, only vigorous activity, unknown [9.9%]), CVD (yes, no), DM (yes, no), SBP (sex specific quartiles), total cholesterol (sex specific quartiles), and when studying the association between respiratory symptoms and mortality, ppFEV₁ (≥100, 80–99, 50–79, <50).

The association between lung function alone or combined with respiratory symptoms and mortality was studied in three different regression models, and the association between respiratory symptoms and mortality was studied in four models (Table 1). First, we explored the association in a simple model (Model 1) adjusting only for age at baseline. In the main model (Model 2) we also adjusted for other potential confounders identified through directed acyclic graphs (DAGs) (34). In additional analyses we adjusted for lung function when studying respiratory symptoms (Model 3), and established risk factors for mortality (Model 4) that could be viewed as possible confounders or mediators although not supported by DAGs (35). Women and men were combined to increase statistical power when studying the association between lung function combined with respiratory symptoms and mortality, and in additional analyses of the association between dyspnoea and CV mortality.

Trend tests across ppFEV₁ levels were calculated using the sex specific median value within each ppFEV₁ level as an ordinal variable in the regression model. For testing the trend across COPD grades we treated the categories as an ordinal variable.

Log-log survival curves for each covariate were inspected and formal tests of interaction with time or log time were produced to evaluate the proportional hazard assumptions. All, but a few, p-values were above 0.05. For all-cause mortality, there were weak evidence for interaction with time for COPD grade 3 or 4 among women (p = 0.049 for time and p = 0.038 for log time). However, the log-log plots showed no clear violation of the proportional hazard assumptions.

In a subsequent sensitivity analysis of the association between respiratory symptoms and all-cause mortality we excluded participants with CVD at baseline. Moreover, all analyses of CV mortality were repeated in a sample where participants with CVD at baseline were included and where CVD was instead adjusted for.

Statistical analyses were conducted using Stata, release 12.1 (StataCorp LP, College Station, Texas). All statistical tests were two-sided.

Ethics

The Norwegian Data Inspectorate licensed the research register (reference 06/00104-39/CGN), and the Regional Committee for Medical Research Ethics (reference 4.2008.59) approved the study. All participants signed informed written consents.

Results

Baseline characteristics

The 10,491 participants we had actual spirometric measurements on, 2655 from the random sample and 7836 from the symptom sample, were weighted to represent 51,854 people from the general population. Table 2 and Table 3 present baseline characteristics according to ppFEV₁ and respiratory symptoms in women and men, respectively. In general, participants with low lung function and participants who reported respiratory symptoms tended to be older, be ever smokers, have less education, be inactive, and have more CVD and DM. BMI was lowest among participants with lowest lung function, and highest among participants reporting respiratory symptoms. Summary statistics of key variables among the symptom, random, and weighted samples are presented in eTable 2. That the percentage or mean of key variables were almost similar in the weighted and the random sample indicates that the weighting was successful.

Table 2.  Baseline characteristics according to lung function and respiratory symptoms among women in the general population (weighted^a)

	Participants	Age [years]	Ever smokers	Education ≥13 years	Body mass index [kg/m^2]	Inactive	Cardiovascular disease	Diabetes mellitus
	%^b(95% CI^c)	Mean(SD)	%^d(95% CI^c)	%^d(95% CI^c)	Mean(SD)	%^d(95% CI^c)	%^d(95% CI^c)	%^d(95% CI^c)
ppFEV1^e ≥100	42.9(40.5–45.3)	49.6(16.7)	42.3(38.5–46.2)	22.4(19.2–25.7)	26.0(3.8)	5.7(3.9–7.6)	2.8(1.6–4.0)	2.0(0.9–3.0)
ppFEV1^e 80–99	44.0(41.6–46.4)	48.1(15.6)	52.3(48.5–56.0)	20.3(17.2–23.3)	26.4(4.7)	6.3(4.5–8.1)	4.5(3.0–5.9)	2.7(1.5–3.9)
ppFEV1^e 50–79	12.2(10.8–13.7)	55.9(15.6)	69.4(63.3–75.4)	11.3(6.9–15.7)	26.9(5.3)	11.7(7.4–16.0)	10.4(6.6–14.2)	3.1(1.1–5.1)
ppFEV1^e <50	0.9(0.6–1.2)	60.9(14.2)	69.0(50.8–87.3)	2.0(0.0–4.0)	24.5(5.5)	32.3(12.0–52.6)	10.0(5.1–14.9)	3.2(0.8–5.6)
No chronic bronchitis	96.6(96.0–3.4)	49.6(16.2)	49.7(47.1–52.3)	20.5(18.4–22.6)	26.2(4.4)	6.4(5.1–7.7)	4.3(3.3–5.2)	2.4(1.7–3.2)
Chronic bronchitis	3.4(2.7–4.0)	53.7(16.3)	58.0(47.9–68.2)	13.8(6.4–21.2)	26.8(5.0)	14.8(8.0–21.5)	11.0(5.1–16.8)	2.6(1.4–3.7)
No wheeze	81.5(80.0–83.0)	49.6(16.3)	46.5(43.6-49.4)	21.1(18.7-23.4)	26.1(4.3)	6.3(4.9-7.7)	3.8(2.7–4.8)	2.4(1.6–3.3)
Wheeze	18.5(17.0–20.0)	50.7(16.0)	66.6(62.6–70.6)	15.6(12.5–18.7)	27.0(4.9)	9.0(6.5–11.4)	7.9(5.7–10.0)	2.7(1.6–3.8)
No dyspnoea	71.9(69.5–74.2)	48.7(15.4)	49.9(46.3–53.4)	22.1(19.2–25.0)	25.5(3.8)	4.4(2.9–5.9)	2.0(1.1–3.0)	1.7(0.8–2.7)
Dyspnoea uphill	11.2(9.5–12.8)	51.1(16.6)	52.3(44.2–60.4)	15.9(10.0–21.8)	27.9(4.7)	12.0(6.4–17.7)	6.9(2.8–10.9)	4.4(1.0–7.7)
Dyspnoea stairs	12.2(10.5–13.9)	54.4(16.9)	52.1(44.6–59.6)	9.8(5.4–14.2)	28.1(5.0)	12.5(7.2–17.8)	12.5(7.6–17.4)	4.5(1.4–7.6)
Dyspnoea walking	3.3(2.6–4.1)	56.6(17.4)	57.6(45.7–69.5)	10.6(1.9–19.3)	30.0(6.3)	17.2(7.5–26.8)	15.5(9.2–21.9)	4.7(2.8–6.6)
Dyspnoea sitting	1.5(1.1–1.9)	49.7(14.4)	77.7(70.0–85.4)	17.2(4.5–29.9)	26.4(5.2)	11.4(6.4–16.5)	6.0(2.9–9.1)	3.0(0.9–5.1)
No symptoms^f	78.4(76.4–80.4)	49.4(15.9)	47.0(43.5–50.5)	20.8(17.9–23.6)	25.9(4.1)	5.9(4.2–7.5)	3.6(2.3–4.8)	2.5(1.5–3.6)
1 symptom	17.5(15.6–19.4)	50.4(15.4)	64.7(59.3–70.2)	16.9(12.7–21.2)	27.1(5.0)	6.2(3.6–8.7)	5.4(3.2–7.7)	1.9(0.8–3.1)
2 symptoms	3.4(2.8–4.0)	53.5(16.3)	58.4(49.5–67.4)	15.0(6.9–23.1)	28.3(5.1)	16.4(7.8–24.9)	13.7(7.8–19.5)	4.4(2.7–6.1)
3 symptoms	0.7(0.5–1.0)	52.3(18.0)	81.0(72.5–89.6)	7.4(2.4–12.4)	27.0(6.5)	21.1(11.2–31.1)	14.8(7.8–21.7)	5.4(1.4–9.3)

^a Data on 1443 women from the random sample and 4076 women from the symptom sample were weighted to represent 28,129 women from the general population.^b Estimated portion of total in the general population.^c CI denotes confidence interval.

^d Estimated prevalence within ppFEV₁ level or respiratory symptom category in the general population, missing excluded.

^e ppFEV₁denotes% predicted forced expiratory volume in 1 second.

^f Symptoms include chronic bronchitis, wheeze, and dyspnoea when walking or sitting.

Table 3.  Baseline characteristics according to lung function and respiratory symptoms among men in the general population (weighted^a)

	Participants	Age [years]	Ever smokers	Education ≥13 years	Body mass index [kg/m^2]	Inactive	Cardiovascular disease	Diabetes mellitus
	%^b(95% CI^c)	Mean(SD)	%^d(95% CI^c)	%^d(95% CI^c)	Mean(SD)	%^d(95% CI^c)	%^d(95% CI^c)	%^d(95% CI^c)
ppFEV1^e ≥100	37.0(34.5–39.6)	48.3(15.4)	50.3(45.7–54.8)	23.2(19.4–27.1)	26.2(3.0)	4.4(2.6–6.3)	5.0(3.1–6.9)	1.9(0.7–3.1)
ppFEV1^e 80–99	47.7(45.1–50.3)	49.4(15.6)	64.8(61.0–68.6)	18.3(15.3–21.4)	26.6(3.6)	8.4(6.2–10.6)	9.2(7.0–11.4)	2.5(1.3–3.7)
ppFEV1^e 50–79	13.7(12.1–15.4)	58.7(15.8)	77.8(72.3–83.3)	12.3(7.9–16.7)	26.5(3.6)	10.2(6.3–14.0)	19.8(14.9–24.7)	3.2(1.5–5.0)
ppFEV1^e <50	1.6(1.2–2.0)	69.7(9.4)	95.0(92.5–97.6)	3.6(1.4–5.7)	25.9(3.9)	18.8(6.2–31.4)	28.3(15.8–40.7)	4.0(1.8–6.2)
No chronic bronchitis	95.7(94.9–96.5)	50.2(16.0)	60.4(57.6–63.1)	19.9(17.6–22.1)	26.4(3.4)	6.9(5.5–8.4)	9.1(7.6–10.6)	2.3(1.5–3.1)
Chronic bronchitis	4.3(3.5–5.1)	56.9(15.4)	81.1(73.0–89.1)	7.5(3.2–11.8)	26.6(4.3)	11.5(6.7–16.3)	13.2(8.7–17.8)	2.5(1.4–3.6)
No wheeze	80.6(78.9–82.3)	50.2(16.0)	57.3(54.1–60.4)	20.3(17.7–22.8)	26.2(3.2)	7.2(5.5–8.8)	8.6(6.9–10.3)	2.0(1.2–2.8)
Wheeze	19.4(17.7–21.1)	52.3(15.9)	80.5(77.2–83.7)	14.4(11.4–17.5)	27.3(3.9)	8.0(5.8–10.2)	12.8(9.9–15.7)	3.8(2.1–5.6)
No dyspnoea	79.4(77.3–81.6)	49.6(15.5)	56.8(53.2–60.4)	21.3(18.3–24.2)	26.0(3.0)	4.8(3.2–6.4)	5.9(4.2–7.5)	1.4(0.6–2.2)
Dyspnoea uphill	9.0(7.5–10.5)	54.5(15.9)	68.9(60.3–77.5)	10.7(5.2–16.2)	27.6(4.3)	6.4(2.4–10.3)	16.6(9.8–23.4)	5.4(1.2–9.6)
Dyspnoea stairs	7.0(5.7–8.4)	57.6(14.9)	82.1(74.4–89.9)	12.8(6.1–19.5)	27.7(4.4)	17.8(9.2–26.4)	20.0(12.3–27.8)	7.3(1.9–12.7)
Dyspnoea walking	3.0(2.1–3.8)	64.0(13.4)	94.5(92.0–97.0)	16.6(3.6–29.6)	28.0(4.4)	26.8(12.2–41.3)	36.6(23.0–50.3)	3.5(1.6–5.3)
Dyspnoea sitting	1.6(1.1–2.0)	57.2(16.6)	82.9(76.8–89.0)	9.6(5.2–13.9)	27.7(4.0)	17.1(11.1–23.1)	22.8(9.9–35.7)	3.3(1.0–5.6)
No symptoms^f	77.0(74.7–79.2)	50.1(15.7)	54.9(51.1–58.7)	21.2(18.1–24.3)	26.1(3.1)	5.5(3.7–7.2)	7.2(5.3–9.1)	1.7(0.8–2.7)
1 symptom	18.4(16.3–20.5)	53.9(15.5)	79.3(74.8–83.8)	15.4(11.4–19.5)	27.2(3.8)	9.3(5.5–13.1)	14.6(10.1–19.1)	3.3(1.2–5.5)
2 symptoms	3.8(3.1–4.6)	57.9(15.3)	88.8(85.7–91.9)	10.1(4.0–16.3)	27.3(4.3)	8.9(6.1–11.7)	19.4(13.0–25.7)	4.0(2.3–5.6)
3 symptoms	0.8(0.5–1.0)	62.6(16.0)	92.2(87.0–97.5)	9.4(3.2–15.6)	27.7(5.0)	28.7(19.4–38.0)	21.4(11.8–30.95)	3.1(0.0–6.2)

^aData on 1212 men from the random sample and 3760 men from the symptom sample were weighted to represent 23,725 men from the general population.

^b Estimated portion of total in the general population.

^c CI denotes confidence interval.

^d Estimated prevalence within ppFEV₁ level or respiratory symptom category in the general population, missing excluded.

^e ppFEV₁denotes% predicted forced expiratory volume in 1 second.

^f Symptoms include chronic bronchitis, wheeze, and dyspnoea when walking or sitting.

Lung function and mortality

The overall all-cause (Table 4) and CV (Table 5) death rates per 1000 person-years were 9.70 and 2.96 in women, and 13.74 and 4.13 in men, respectively. Compared to participants with ppFEV₁ ≥100 or normal airflow, women and men with ppFEV₁ <80 or COPD grade 2 or higher had increased all-cause mortality (Table4). For every 10% decrease in ppFEV₁ the adjusted HRs for all-cause mortality was 1.17 (95% CI 1.09–1.25) in women and 1.23 (95% CI 1.16–1.30) in men. Additional adjustments for BMI, physical activity, CVD, DM, SBP, and total cholesterol did not materially change the results (eTable 3).

Table 4.  Death rates and adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) for the associations between lung function and all-cause mortality (weighted)

					Model 1: Age	Model 2: Model 1 + smoking, education
Lung function	Person-years	Deaths	Death rate^a	(95% CI)	HR^bPtrend^b	HR^C(95% CI)^CPtrend^C
Women	361,010	3501	9.70	(8.52–11.08)
ppFEV1^d ≥100	156,685	1285	8.20	(6.55–10.40)	1.00	1.00(Reference)
ppFEV1^d 80–99	160,258	1214	7.58	(6.06–9.60)	1.20	1.17(0.85–1.61)
ppFEV1^d 50–79	41,850	847	20.24	(15.91–26.06)	1.90	1.73(1.19–2.51)
ppFEV1^d <50	2217	155	69.92	(44.06–111.32)	7.61<0.001	6.85 4.46–10.52)<0.001
Normal airflow^e	315,850	2370	7.50	(6.37–8.89)	1.00	1.00(Reference)
COPD grade 1^f	9828	227	23.10	(14.34–39.53)	1.36	1.42(0.82–2.43)
COPD grade 2^g	17,805	512	28.76	(21.25–39.55)	2.30	2.02(1.39–2.95)
COPD grade 3 or 4^h	2194	153	69.75	(43.71–111.68)	7.20<0.001	6.50(4.33–9.75)<0.001
ppFEV1^d 10% decrease					1.19	1.17(1.09–1.25)
Men	295,762	4064	13.74	(12.20–15.52)
ppFEV1^d ≥100`	114,302	929	8.13	(6.25–10.76)	1.00	1.00(Reference)
ppFEV1^d 80–99	142,332	1657	11.64	(9.58–14.27)	1.30	1.26(0.90–1.77)
ppFEV1^d 50–79	36,210	1192	32.92	(26.97–40.44)	2.03	1.91(1.32–2.75)
ppFEV1^d <50	2919	286	97.99	(74.09–130.28)	4.00<0.001	3.88(2.60–5.79)<0.001
Normal airflow^e	244,721	2382	9.73	(8.27–11.53)	1.00	1.00(Reference)
COPD grade 1^f	17,291	381	22.03	(14.89–33.71)	1.27	1.23(0.81–1.85)
COPD grade 2^g	22,224	880	39.60	(31.45–50.21)	1.73	1.63(1.17–2.29)
COPD grade 3 or 4^h	2410	260	107.86	(88.04–134.17)	3.75<0.001	3.57(2.60–4.91)<0.001
ppFEV1 10% decrease					1.23	1.23(1.16–1.30)

^a Per 1000 person-years.

^b Model 1: Adjusted for age (<40, 40–49, …, ≥80 years).

^c Model 2: Adjusted for all in Model 1 + smoking (never, former, current, unknown), and education (<10, 10–12, ≥13 years, unknown)

^d ppFEV₁denotes% predicted forced expiratory volume in 1 second.

^e Normal airflow defined as forced expiratory volume in 1 second (FEV₁)/forced vital capacity (FVC) ≥70% and% predicted FVC ≥80. Participants with restrictive lung function impairment (FEV₁/FVC ≥ 70% and ppFVC < 80) were excluded from the COPD analyses.

^f COPD grade 1 denotes chronic obstructive pulmonary disease defined as FEV₁/FVC < 70% and ppFEV₁≥80.

^g COPD grade 2 defined as FEV₁/FVC < 70% and 50 ≤ ppFEV₁< 80.

^h COPD grade 3 or 4 defined as FEV₁/FVC < 70% and ppFEV₁< 50.

Table 5.  Death rates and adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) for the associations between lung function and cardiovascular mortality (weighted) in participants without CVD at baseline

					Model 1: Age	Model 1 + Model 2: smoking, education
Lung function	Person-years	Deaths	Death rate^a	(95% CI)^a	HR^bPtrend^b	HR^C(95% CI)^CPtrend^C
Women	346,952	1028	2.96	(2.32–3.84)
ppFEV1^d ≥100	152,857	524	3.43	(2.39–5.10)	1.00	1.00(Reference)
ppFEV1^d 80–99	153,886	287	1.87	(1.18–3.14)	0.77	0.73(0.40–1.35)
ppFEV1^d 50–79	38,171	198	5.19	(3.20–8.96)	1.39	1.09(0.56–2.13)
ppFEV1^d <50	2037	19	9.33	(5.11–17.19)	2.57 0.607	2.21(1.14–4.29)0.932
Normal airflow^e	304,188	784	2.58	(1.93–3.53)	1.00	1.00(Reference)
COPD grade 1^f	9464	50	5.28	(1.64–25.93)	0.70	0.72(0.23–2.24)
COPD grade 2^g	15,953	94	5.89	(3.13–12.26)	1.42	1.20(0.58–2.50)
COPD grade 3 or 4^h	2014	19	9.44	(5.15–17.43)	2.74 0.270	2.59(1.37–4.88)0.505
ppFEV1^d 10% decrease					1.07	1.03(0.91–1.16)
Men	273,482	1130	4.13	(3.28–5.28)
ppFEV1^d ≥100	109,247	220	2.01	(1.17–3.79)	1.00	1.00(Reference)
ppFEV1^d 80–99	131,867	476	3.61	(2.50–5.40)	1.72	1.69(0.83–3.42)
ppFEV1^d 50–79	30,114	373	12.39	(8.44–18.85)	2.83	2.70(1.21–6.02)
ppFEV1^d <50	2255	61	27.06	(13.26–62.59)	3.82 0.001	3.86(1.51–9.89)0.005
Normal airflow ^e	228,829	539	2.36	(1.67–3.44)	1.00	1.00(Reference)
COPD grade 1^f	15,109	121	8.01	(3.92–18.84)	1.78	1.66(0.70–3.96)
COPD grade 2^g	18,502	313	16.92	(11.09–26.92)	2.46	2.35(1.20–4.60)
COPD grade 3 or 4^h	1926	60	31.15	(16.08–69.81)	3.46 <0.001	3.34(1.49–7.50)0.002
ppFEV1^d 10% decrease					1.24	1.24(1.10–1.39)

^a Per 1000 person-years.

^b Model 1: Adjusted for age (<40, 40–49,  .  .  .  , ≥80 years).

^c Model 2: Adjusted for all in Model 1 + smoking (never, former, current, unknown), and education (<10, 10–12, ≥13 years, unknown).

^d ppFEV₁denotes% predicted forced expiratory volume in 1 second.

^e Normal airflow defined as forced expiratory volume in 1 second (FEV₁)/forced vital capacity (FVC) ≥ 70% and% predicted FVC ≥80. Participants with restrictive lung function impairment (FEV₁/FVC ≥ 70% and ppFVC < 80) were excluded from the COPD analyses.

^f COPD grade 1 denotes Global Initiative for Chronic Obstructive Lung Disease airflow limitation defined as FEV₁/FVC < 70% and ppFEV₁≥ 80.

^g COPD grade 2 defined as FEV₁/FVC < 70% and 50 ≤ ppFEV₁< 80.

^h COPD grade 3 or 4 defined as FEV₁/FVC < 70% and ppFEV₁< 50.

CV mortality was increased in women with ppFEV₁ <50 or COPD grade 3 or 4 and in men with ppFEV₁ <80 or COPD grade 2 or higher (Table 5). For every 10% decrease in ppFEV₁ the adjusted HRs for CV mortality were 1.03 (95% CI 0.91–1.16) in women and 1.24 (95% CI 1.10–1.39) in men. Additional adjustments for BMI, physical activity, DM, SBP, and total cholesterol (eTable4) or including participants with CVD at baseline in a sensitivity analysis (eTable 5) did not materially change the results.

Respiratory symptoms and mortality

Chronic bronchitis, dyspnoea when walking, dyspnoea when sitting (among women only), and number of respiratory symptoms were positively associated with all-cause mortality in models not adjusted for lung function (Table 6). However, only dyspnoea when walking remained positively associated with all-cause mortality when adjusting for lung function (HR 1.73 [95% CI 1.04–2.89] in women and HR 1.57 [95% CI 1.04–2.36] in men). With additional adjustments for DM, SBP, and total cholesterol, the association between dyspnoea when walking and all-cause mortality in women attenuated (eTable 6). When excluding participants with CVD at baseline in a sensitivity analysis none of the respiratory symptoms were associated with all-cause mortality when adjusting for lung function (eTable 7). Within lung function levels, participants reporting dyspnoea when walking or sitting had generally higher HRs for all-cause mortality than participants without these symptoms (Figure 2). Similar trends were not seen for chronic bronchitis or wheeze within lung function levels (eFigure 1a and 1b). Additional adjustments for DM, SBP, and total cholesterol did not materially change the results (data not shown).

Figure 2.  Association of the combined exposure of % predicted forced expiratory volume in 1 second (ppFEV₁) and reporting of dyspnoea when walking or sitting with all-cause mortality. The reference category consists of people with ppFEV₁ ≥100 not reporting dyspnoea when walking or sitting. The bars represent 95% confidence intervals. Adjusted for age (<40, 40–49, …, ≥80 years), sex (woman, man), smoking (never, former, current, unknown), education (<10, 10–12, ≥13 years, unknown), body mass index (<18.5, 18.5–24.9, 25.0–29.9, ≥30.0 kg/m²), physical activity (inactive, light activity <1 hours/week, light activity 1–2 hours/week, light activity ≥3 hours/week, only vigorous activity, unknown), and cardiovascular disease (yes, no).

Table 6.  Death rates and adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) for the associations between respiratory symptoms and all-cause mortality (weighted)

					Model 1:Age	Model 2:Model 1 + smoking, education, BMI, physical activity, CVD	Model 3:Model 2 + lung function
Respiratory symptom	Person-years	Deaths	Death rate^a	(95% CI)	HR^b	HR^c(95% CI)^c	HR^d(95% CI)^d
Women
No chronic bronchitis	344,075	3191	9.27	(8.08–10.70)	1.00	1.00(Reference)	1.00(Reference)
Chronic bronchitis	11,588	218	18.81	(12.84–28.59)	1.70	1.52(1.03–2.23)	1.27(0.85–1.89)
No wheeze	294,178	2659	9.04	(7.72–10.65)	1.00	1.00(Reference)	1.00(Reference)
Wheeze	65,154	796	12.22	(10.00–15.06)	1.42	1.21(0.91–1.60)	1.05(0.79–1.40)
No dyspnoea	191,547	1337	6.98	(5.58–8.84)	1.00	1.00(Reference)	1.00(Reference)
Dyspnoea uphill	28,968	368	12.70	(8.71–19.23)	1.52	1.48(0.94–2.34)	1.34(0.83–2.15)
Dyspnoea stairs	31,380	401	12.78	(9.12–18.37)	1.17	1.07(0.66–1.74)	0.97(0.59–1.59)
Dyspnoea walking	7998	231	28.88	(20.07–42.68)	2.62	2.43(1.51–3.92)	1.73(1.04–2.89)
Dyspnoea sitting	3755	61	16.24	(8.22–36.46)	3.06	2.51(1.26–5.01)	1.94(0.92–4.09)
No symptoms^e	203,968	1615	7.92	(6.45–9.82)	1.00	1.00(Reference)	1.00(Reference)
1 symptom	44,629	519	11.63	(8.74–15.79)	1.54	1.38(0.96–2.00)	1.29(0.90–1.86)
2 symptoms	8318	147	17.67	(12.85–24.75)	2.17	1.95(1.38–2.76)	1.36(0.89–2.07)
3 symptoms	1770	38	21.47	(13.76–33.36)	2.61	1.75(1.08–2.84)	1.16(0.72–1.86)
Men
No chronic bronchitis	278,849	3623	12.99	(11.42–14.84)	1.00	1.00(Reference)	1.00(Reference)
Chronic bronchitis	11,398	318	27.90	(20.62–38.20)	1.63	1.49(1.10–2.02)	1.16(0.88–1.55)
No wheeze	238,684	3057	12.81	(11.07–14.90)	1.00	1.00(Reference)	1.00(Reference)
Wheeze	55,582	1003	18.05	(15.11–21.68)	1.32	1.20(0.93–1.55)	0.97(0.74–1.26)
No dyspnoea	176,705	1953	11.05	(9.21–13.37)	1.00	1.00(Reference)	1.00(Reference)
Dyspnoea uphill	19,322	379	19.61	(14.11–27.93)	1.32	1.27(0.88–1.84)	1.02(0.70–1.47)
Dyspnoea stairs	14,417	335	23.24	(16.53–33.33)	1.12	0.91(0.56–1.46)	0.76(0.48–1.22)
Dyspnoea walking	4756	281	59.08	(39.43–89.71)	2.66	2.41(1.59–3.65)	1.57(1.04–2.36)
Dyspnoea sitting	3222	103	31.97	(21.06–52.08)	1.62	1.51(0.83–2.76)	1.22(0.68–2.17)
No symptoms^e	166,163	1978	11.90	(9.92–14.40)	1.00	1.00(Reference)	1.00(Reference)
1 symptom	37,821	679	17.95	(13.94–23.44)	1.21	1.14(0.81–1.62)	0.98(0.70–1.38)
2 symptoms	7402	256	34.59	(26.08–46.75)	2.03	1.82(1.29–2.57)	1.25(0.88–1.79)
3 symptoms	1194	70	58.63	(34.27–110.65)	3.17	2.59(1.64–4.09)	1.48(0.95–2.30)

^a Per 1000 person-years.

^b Model 1: Adjusted for age (<40, 40–49,  .  .  .  , ≥80 years).

^c Model 2: Adjusted for all in Model 1 + smoking (never, former, current, unknown), education (<10, 10–12, ≥13 years, unknown), body mass index (<18.5, 18.5–24.9, 25.0–29.9, ≥30.0 kg/m²), physical activity (inactive, light activity <1 hours/week, light activity 1–2 hours/week, light activity ≥3 hours/week, only vigorous activity, unknown), and cardiovascular disease (yes, no).

^d Model 3: Adjusted for all in Model 2 +% predicted forced expiratory volume in 1 second (≥100, 80–99, 50–79, <50).

^e Symptoms include chronic bronchitis, wheeze, and dyspnoea when walking or sitting.

None of the respiratory symptoms were positively associated with CV mortality when adjusting for lung function (eTable 8). The results did not materially change with additionally adjustment for DM, SBP, and total cholesterol (eTable 9) or when including participants with CVD at baseline in a sensitivity analysis (eTable 10). Participants reporting respiratory symptoms did not have a consistent trend for increased CV mortality compared to participants not reporting respiratory symptoms within lung function levels (data not shown). However, when combining women and men, dyspnoea when walking was positively associated with CV mortality when participants with CVD at baseline were included in the model (HR 1.67 [95% CI 1.05-2.66]) (eTable 11).

There were little differences in baseline characteristics (eTable 12) and HRs for all-cause and CV mortality (eTable 13) between participants with and without data on the dyspnoea scale, indicating no substantial selection bias due to missing data.

Discussion

In this large epidemiological study from a general population in Norway we found pre-bronchodilator lung function to be strongly and inversely associated with all-cause and CV mortality. Although chronic bronchitis, dyspnoea when walking or sitting, and number of respiratory symptoms were positively associated with all-cause mortality in crude models, only dyspnoea when walking remained positively associated with all-cause mortality when controlling for lung function. None of the respiratory symptoms were associated with CV mortality independent of lung function in sex specific models.

In accordance with findings from previous epidemiological studies (4–19), we found increased all-cause and CV mortality with lower lung function. However, none of these previous studies have classified participants according to both ppFEV₁ and COPD grades, and few have done sex specific analyses. For the main analyses of CV mortality we excluded participants with CVD at baseline to avoid reverse causation. Thus, our results indicate that impaired lung function may be causally associated with CV mortality. Possible mechanisms that may explain our findings include systemic inflammation, impaired functional capacity, muscle dysfunction, malnutrition, oxidative stress, and comorbidities such as CVD, depression, lung cancer, and DM (36).

In a population-based study from the United States, reporting at least one of cough, phlegm, wheeze, or breathlessness increased the HR for all-cause mortality compared to not reporting any symptom within pre-bronchodilator levels of COPD (12). However, there was no investigation of which respiratory symptoms were responsible for the association with mortality, and the analyses were not conducted sex specific. In the current and in other studies (15, 19), dyspnoea, but not chronic bronchitis or wheeze, has been found to be associated with all-cause mortality independent of lung function.

Interestingly, among the four levels of dyspnoea only dyspnoea when walking was associated with all-cause mortality in the current study. However, none of the respiratory symptoms remained associated with all-cause mortality independent of lung function when we excluded participants with CVD at baseline, indicating that CVD could explain some of the association between dyspnoea when walking and mortality. Nevertheless, excluding people also results in reduced power to detect an association. Future studies aiming at scrutinising the complex mechanisms involved in the sensation of dyspnoea (37) may be needed to explain our findings.

Somewhat unexpectedly, we found no association between dyspnoea and CV mortality independent of lung function in sex specific models. This lack of association may be due to low statistical power, misclassification of CV deaths, residual confounding, or it may reflect the reality. Additional analyses combining women and men increased the statistical power, but only the association between dyspnoea when walking and CV mortality was statistically significant when participants with CVD at baseline were included. Others have found breathlessness to be associated with CV mortality in 40–64 years old men after controlling for FEV₁ (18, 19). Although these studies did not exclude men with CVD at baseline, they adjusted for baseline myocardial ischemia. A Dutch study with over 40 years of follow-up found dyspnoea to be clearly associated with CV mortality also after adjusting for lung function (24). However, CVD at baseline was not accounted for, the analyses were not conducted sex specific, and CV death was coded differently than in the current study (24).

Strengths of the current study include the large population-based sample size which gave us the possibility to do sex specific analyses, to exclude participants with CVD at baseline, and to study lung function and respiratory symptoms as a combined exposure. In addition, our participants had a large span in age and lung function, and they were included independent of clinical diagnoses. HUNT2 had a relatively high response rate, and a non-responder study did not indicate serious selection bias (38).

All-cause mortality is a robust outcome since the Cause of Death Registry in Norway is practically complete and there is minimal room for misclassification. However, the results on CV mortality must be interpreted with caution since it is possible that CV deaths may have been registered as deaths of other causes and vice versa. If this possible misclassification is non-differential it may dilute the real effect estimates (35), which could explain the lack of association between respiratory symptoms and CV mortality in our data. About one quarter of the participants in the HUNT2 Lung Study did not return the lung specific questionnaire, which included the dyspnoea scale. However, it is unlikely that missing data on the dyspnoea scale have biased the results since there were no large differences in baseline characteristics and HRs for all-cause and CV mortality between participants with and without data on the dyspnoea scale. Despite thorough adjustment for possible confounding factors, there may still be residual confounding due to poorly measured or unmeasured variables. Additionally, we were not able to account for possible changes in factors such as smoking, physical activity, and BMI during follow-up. Finally, respiratory symptoms were self-reported and not validated.

In summary, pre-bronchodilator lung function was strongly and inversely associated with all-cause and CV mortality, and dyspnoea when walking was positively associated with all-cause mortality independent of lung function. Chronic bronchitis, dyspnoea when sitting, and number of respiratory symptoms were positively associated with all-cause mortality only in models where lung function was not controlled for. Respiratory symptoms were not associated with CV mortality independent of lung function. Our results suggest that objectively measured lung function is superior to subjectively reported respiratory symptoms in mortality studies.

Declaration of Interest Statement

The authors report no conflicts of interest. This study was financially supported by the Norwegian ExtraFoundation for Health and Rehabilitation through EXTRA funds (project 2007/2/0214); the Liaison Committee between the Central Norway Regional Health Authority and the Norwegian University of Science and Technology (project 46041600); and the Leif Richard Erichsen and wife Maren Hertzberg Erichsen's fund for Norwegian medical research. The HUNT2 Lung Study received funding without obligation from AstraZeneca.

The authors alone are responsible for the content and writing of the paper.

Acknowledgements

The second survey of the Nord-Trøndelag Health Study 1995–1997 (HUNT2) was a collaboration between HUNT Research Centre (Faculty of Medicine, Norwegian University of Science and Technology), Nord-Trøndelag County Council, Central Norway Regional Health Authority, and the Norwegian Institute of Public Health.

WEB APPENDIX

Lung Function and Respiratory Symptoms in Association with Mortality: The HUNT Study

L. Leivseth, T.I.L. Nilsen, X.M. Mai, R. Johnsen, and A. Langhammer

eTable 1. The dyspnoea scale used in the Nord-Trøndelag Health Study 1995–1997 Lung Study

Table

Measuring	Wording
Dyspnoea walking uphill	Do you become more short of breath than people your age when walking uphill?
Dyspnoea climbing stairs	Do you become short of breath when you climb two flights of stairs at a normal pace?
Dyspnoea walking	Do you become short of breath when walking on flat ground at a normal pace?
Dyspnoea sitting	Are you short of breath when sitting still?

eTable 2. Respiratory symptoms, lung function, medical diagnoses, and other characteristics of participants in the symptom, random, and weighted sample

Table

	Women			Men
	Symptomn = 4076, 52.0%^a	Randomn = 1443, 54.4%^a	Weighted n = 28,129, 54.2%^a	Symptom n = 3760, 48.0%^a	Random n = 1212, 45.6%^a	Weighted n = 23,725, 45.8%^a
Respiratory symptoms, n (%) ^b
Chronic bronchitis	513 (13.4)	46 (3.2)	939 (3.4)	540 (15.1)	60 (5.0)	999 (4.3)
Wheeze	2669 (65.8)	291 (20.3)	5183 (18.5)	2375 (63.4)	245 (20.3)	4577 (19.4)
Dyspnoea uphill	620 (20.4)	118 (11.1)	2296 (11.2)	582 (20.7)	85 (9.3)	1579 (9.0)
Dyspnoea stairs	772 (25.4)	134 (12.6)	2502 (12.2)	458 (16.3)	71 (7.8)	1232 (7.0)
Dyspnoea walking	361 (11.9)	36 (3.4)	682 (3.3)	242 (8.6)	30 (3.3)	519 (3.0)
Dyspnoea sitting	208 (6.8)	16 (1.5)	300 (1.5)	204 (7.3)	11 (1.2)	272 (1.6)
Lung function, mean (SD)
FEV1^c [L]	2.6 (0.8)	2.8 (0.7)	2.8 (0.7)	3.4 (1.1)	3.8 (1.0)	3.8 (1.0)
ppFEV1^d	88.7 (18.9)	96.8 (15.9)	96.7 (16.1)	84.9 (20.6)	94.5 (16.1)	94.4 (16.5)
FVC^e	3.3 (0.8)	3.5 (0.8)	3.5 (0.8)	4.6 (1.2)	4.9 (1.0)	4.9 (1.0)
FEV1/FVC [%]	76.6 (10.3)	80.0 (7.2)	79.9 (7.4)	72.6 (11.8)	77.0 (8.5)	76.9 (8.7)
Lung function, n (%) a
ppFEV1^d ≥ 100	1132 (27.8)	621 (43.0)	12,070 (42.9)	859 (22.9)	443 (36.6)	8788 (37.0)
ppFEV1^d 80–99	1814 (44.5)	638 (44.2)	12,370 (44.0)	1599 (42.5)	591 (48.8)	11,310 (47.7)
ppFEV1^d 50–79	968 (23.8)	172 (11.9)	3439 (12.2)	1039 (27.6)	161 (13.3)	3252 (13.7)
ppFEV1^d < 50	162 (4.0)	12 (0.8)	250 (0.9)	263 (7.0)	17 (1.4)	375 (1.6)
Normal airflow^f	2971 (72.9)	1257 (87.1)	24,330 (86.5)	2404 (63.9)	978 (80.7)	19,152 (80.7)
COPD grade 1^g	153 (3.8)	39 (2.7)	781 (2.8)	235 (6.3)	76 (6.3)	1451 (6.1)
COPD grade 2^h	521 (12.8)	74 (5.1)	1545 (5.5)	678 (18.0)	104 (8.6)	2074 (8.7)
COPD grade 3 or 4^i	159 (3.9)	12 (0.8)	247 (0.9)	256 (6.8)	15 (1.2)	328 (1.4)
Medical diagnoses (ever had), n (%)^b
Myocardial infarction	91 (2.2)	17 (1.2)	317 (1.1)	259 (6.9)	64 (5.3)	1254 (5.3)
Angina pectoris	237 (5.8)	35 (2.4)	709 (2.5)	362 (9.6)	59 (4.9)	1238 (5.2)
Stroke or cerebral haemorrhage	80 (2.0)	27 (1.9)	468 (1.7)	88 (2.3)	21 (1.7)	432 (1.8)
Diabetes mellitus	134 (3.3)	41 (2.8)	688 (2.4)	148 (3.9)	27 (2.2)	574 (2.4)
Asthma	1519 (51.4)	93 (8.9)	1631 (8.0)	1339 (48.6)	75 (8.3)	1452 (8.4)
Chronic bronchitis or emphysema	490 (16.8)	33 (3.2)	694 (3.5)	533 (19.7)	48 (5.3)	866 (5.0)
Other, mean (SD)
Age [years]	50.3 (17.0)	49.8 (16.4)	49.8 (16.3)	51.8 (16.8)	50.5 (16.0)	50.6 (16.0)
Body mass index [kg/m^2]	27.2 (5.1)	26.2 (4.4)	26.3 (4.4)	27.0 (3.9)	26.5 (3.4)	26.4 (3.4)
Systolic blood pressure [mmHg]	135.8 (22.7)	135.3 (23.7)	135.6 (23.9)	140.6 (19.3)	140.7 (19.2)	140.6 (19.3)
Total cholesterol [mmol/l]	6.0 (1.3)	5.9 (1.3)	6.0 (1.3)	5.9 (1.2)	5.9 (1.1)	5.9 (1.1)
Other, n (%)^b
Education ≥ 13 years	648 (16.8)	271 (19.8)	5346 (20.0)	562 (15.7)	222 (19.1)	4337 (19.1)
Ever smoker	2280 (59.0)	697 (51.1)	13,351 (50.2)	2592 (71.7)	701 (61.3)	13,819 (61.7)
Inactive	360 (10.2)	87 (6.8)	1701 (6.8)	321 (9.5)	83 (7.4)	1582 (7.3)

^a Percent within each sample.

^b Percent within variable, sample, and sex, missing excluded.

^c FEV₁ denotes forced expiratory volume in one second.

^d ppFEV₁ denotes% predicted FEV₁.

^e FVC denotes forced vital capacity.

^f Normal airflow defined by FEV₁/FVC ≥ 70% and% predicted FVC ≥ 80.

^g COPD Grade 1 denotes chronic obstructive pulmonary disease defined as FEV₁/FVC <70% and ppFEV₁ ≥ 80.

^h COPD Grade 2 defined as FEV₁/FVC < 70% and 80 > ppFEV₁ ≥ 50.

ⁱ COPD Grade 3 or 4 defined as FEV₁/FVC < 70% and ppFEV₁< 50.

eTable 3. Adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) for the associations between lung function and all-cause mortality (weighted)

Table

	Model 1:Age		Model 2:Model 1 + smoking, education		Model 4:Model 2 + BMI, physical activity, CVD, DM, SBP, total cholesterol
Lung function	HR^a (95% CI)^a	Ptrend^a	HR^b (95% CI)^b	Ptrend^b	HR^c (95% CI)^c	Ptrend^c
Women
ppFEV1^d ≥ 100	1.00 (Reference)		1.00 (Reference)		1.00 (Reference)
ppFEV1^d 80–99	1.20 (0.87–1.65)		1.17 (0.85–1.61)		1.11 (0.80–1.56)
ppFEV1^d 50–79	1.90 (1.33–2.72)		1.73 (1.19–2.51)		1.66 (1.15–2.40)
ppFEV1^d < 50	7.61 (5.04–11.50)	<0.001	6.85 (4.46–10.52)	<0.001	5.67 (3.61–8.92)	0.001
Normal airflow^e	1.00 (Reference)		1.00 (Reference)		1.00 (Reference)
COPD grade 1^f	1.36 (0.76–2.43)		1.42 (0.82–2.43)		1.55 (0.90–2.65)
COPD grade 2^g	2.30 (1.57–3.35)		2.02 (1.39–2.95)		1.97 (1.35–2.89)
COPD grade 3 or 4^h	7.20 (4.84–10.71)	<0.001	6.50 (4.33–9.75)	<0.001	5.70 (3.69–8.81)	<0.001
ppFEV1^d 10% decrease	1.19 (1.11–1.28)		1.17 (1.09–1.25)		1.15 (1.08–1.24)
Men
ppFEV1^d ≥ 100	1.00 (Reference)		1.00 (Reference)		1.00 (Reference)
ppFEV1^d 80–99	1.30 (0.94–1.81)		1.26 (0.90–1.77)		1.12 (0.80–1.58)
ppFEV1^d 50–79	2.03 (1.45–2.85)		1.91 (1.32–2.75)		1.66 (1.15–2.37)
ppFEV1^d < 50	4.00 (2.75–5.82)	<0.001	3.88 (2.60–5.79)	<0.001	3.29 (2.22–4.87)	<0.001
Normal airflow^e	1.00 (Reference)		1.00 (Reference)		1.00 (Reference)
COPD grade 1^f	1.27 (0.84–1.92)		1.23 (0.81–1.85)		1.28 (0.84–1.93)
COPD grade 2^g	1.73 (1.26–2.37)		1.63 (1.17–2.29)		1.58 (1.13–2.21)
COPD grade 3 or 4^h	3.75 (2.79–5.05)	<0.001	3.57 (2.60–4.91)	<0.001	3.36 (2.43–4.63)	<0.001
ppFEV1^d 10% decrease	1.23 (1.17–1.30)		1.23 (1.16–1.30)		1.20 (1.14–1.27)

^a Model 1: Adjusted for age (<40, 40–49,  .  .  .  , ≥ 80 years).

^b Model 2: Adjusted for all in Model 1 + smoking (never, former, current, unknown), and education (<10, 10–12, ≥ 13 years, unknown).

^c Model 4: Adjusted for all in Model 2 + body mass index [BMI] (<18.5, 18.5–24.9, 25.0–29.9, ≥ 30.0 kg/m²), physical activity (inactive, light activity <1 hours/week, light activity 1–2 hours/week, light activity ≥ 3 hours/week, only vigorous activity, unknown), cardiovascular disease [CVD] (yes, no), diabetes mellitus [DM] (no, yes), systolic blood pressure [SBP] (sex specific quartiles), and total cholesterol (sex specific quartiles).

^d ppFEV1 denotes% predicted forced expiratory volume in 1 second.

^e Normal airflow defined as forced expiratory volume in 1 second (FEV₁)/forced vital capacity (FVC) ≥ 70% and% predicted FVC ≥ 80. Participants with restrictive lung function impairment (FEV₁/FVC ≥ 70% and ppFVC < 80) were excluded from the COPD analyses.

^f COPD grade 1 denotes chronic obstructive pulmonary disease defined as FEV₁/FVC <70% and ppFEV1 ≥80.

^g COPD grade 2 defined as FEV₁/FVC <70% and 50 ≤ ppFEV₁ <80.

^h COPD grade 3 or 4 defined as FEV₁/FVC <70% and ppFEV₁ <50.

eTable 4. Adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) for the associations between lung function and cardiovascular mortality (weighted) in participants without CVD at baseline

Table

	Model 1:Age		Model 2:Model 1 + smoking, education		Model 4:Model 2 + BMI, physical activity, DM, SBP, total cholesterol
Lung function	HR^a (95% CI)^a	Ptrend^a	HR^b (95% CI)^b	Ptrend^b	HR^c (95% CI)^c	Ptrend^c
Women
ppFEV1^d ≥ 100	1.00 (Reference)		1.00 (Reference)		1.00 (Reference)
ppFEV1^d 80–99	0.77 (0.42–1.41)		0.73 (0.40–1.35)		0.76 (0.39–1.47)
ppFEV1^d 50–79	1.39 (0.75–2.59)		1.09 (0.56–2.13)		1.09 (0.56–2.11)
ppFEV1^d < 50	2.57 (1.37–4.84)	0.607	2.21 (1.14–4.29)	0.932	2.36 (1.08–5.19)	0.995
Normal airflow^e	1.00 (Reference)		1.00 (Reference)		1.00 (Reference)
COPD grade 1^f	0.70 (0.23–2.14)		0.72 (0.23–2.24)		0.65 (0.19–2.21)
COPD grade 2^g	1.42 (0.71–2.83)		1.20 (0.58–2.50)		1.09 (0.52–2.28)
COPD grade 3 or 4^h	2.74 (1.49–5.05)	0.270	2.59 (1.37–4.88)	0.505	2.86 (1.32–6.21)	0.698
ppFEV1^d 10% decrease	1.07 (0.95–1.21)		1.03 (0.91–1.16)		1.02 (0.90–1.16)
Men
ppFEV1^d ≥ 100	1.00 (Reference)		1.00 (Reference)		1.00 (Reference)
ppFEV1^d 80–99	1.72 (0.88–3.35)		1.69 (0.83–3.42)		1.57 (0.80–3.07)
ppFEV1^d 50–79	2.83 (1.38–5.79)		2.70 (1.21–6.02)		2.54 (1.15–5.59)
ppFEV1^d < 50	3.82 (1.56–9.39)	0.001	3.86 (1.51–9.89)	0.005	3.06 (1.18–7.90)	0.010
Normal airflow^e	1.00 (Reference)		1.00 (Reference)		1.00 (Reference)
COPD grade 1^f	1.78 (0.81–3.92)		1.66 (0.70–3.96)		2.09 (0.92–4.77)
COPD grade 2^g	2.46 (1.34–4.53)		2.35 (1.20–4.60)		2.41 (1.17–4.97)
COPD grade 3 or 4^h	3.46 (1.58–7.58)	<0.001	3.34 (1.49–7.50)	0.002	3.26 (1.42–7.50)	0.002
ppFEV1^d 10% decrease	1.24 (1.12–1.38)		1.24 (1.10–1.39)		1.22 (1.09–1.38)

^a Model 1: Adjusted for age (<40, 40–49,  .  .  .  , ≥ 80 years).

^b Model 2: Adjusted for all in Model 1 + smoking (never, former, current, unknown), and education (<10, 10–12, ≥13 years, unknown).

^c Model 4: Adjusted for all in Model 2 + body mass index [BMI] (<18.5, 18.5–24.9, 25.0–29.9, ≥30.0 kg/m²), physical activity (inactive, light activity <1 hours/week, light activity 1–2 hours/week, light activity ≥ 3 hours/week, only vigorous activity, unknown), diabetes mellitus [DM] (no, yes), systolic blood pressure [SBP] (sex specific quartiles), and total cholesterol (sex specific quartiles).

^d ppFEV₁ denotes% predicted forced expiratory volume in 1 second.

^e Normal airflow defined as forced expiratory volume in 1 second (FEV₁)/forced vital capacity (FVC) ≥ 70% and% predicted FVC ≥80. Participants with restrictive lung function impairment (FEV₁/FVC ≥ 70% and ppFVC <80) were excluded from the COPD analyses.

^f COPD grade 1 denotes chronic obstructive pulmonary disease defined as FEV₁/FVC <70% and ppFEV₁ ≥ 80.

^g COPD grade 2 defined as FEV₁/FVC <70% and 50≤ ppFEV₁ <80.

^h COPD grade 3 or 4 defined as FEV₁/FVC <70% and ppFEV₁ <50.

eTable 5. Sensitivity analyses –participants with CVD at baseline included: Adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) for the associations between lung function and cardiovascular mortality (weighted)

Table

	Model 1:Age		Model 2:Model 1 + smoking, education		Model 4:Model 2 + BMI, physical activity, CVD, DM, SBP, total cholesterol
Lung function	HR^a (95% CI)^a	Ptrend^a	HR^b (95% CI)^b	Ptrend^b	HR^c (95% CI)^c	Ptrend^c
Women
ppFEV1^d ≥100	1.00 (Reference)		1.00 (Reference)		1.00 (Reference)
ppFEV1^d 80–99	0.97 (0.58–1.61)		0.96 (0.58–1.59)		0.90 (0.52–1.55)
ppFEV1^d 50–79	1.47 (0.85–2.53)		1.28 (0.71–2.29)		1.14 (0.63–2.08)
ppFEV1^d <50	2.79 (1.64–4.76)	0.207	2.43 (1.39–4.25)	0.422	2.07 (1.09–3.94)	0.689
Normal airflow^e	1.00 (Reference)		1.00 (Reference)		1.00 (Reference)
COPD grade 1^f	0.61 (0.22–1.72)		0.65 (0.23–1.85)		0.71 (0.23–2.13)
COPD grade 2^g	1.47 (0.82–2.65)		1.32 (0.71–2.47)		1.27 (0.68–2.39)
COPD grade 3 or 4^h	2.70 (1.63–4.46)	0.164	2.46 (1.46–4.13)	0.305	2.36 (1.27–4.36)	0.362
ppFEV1^d 10% decrease	1.10 (1.00–1.22)		1.07 (0.96–1.19)		1.05 (0.94–1.17)
Men
ppFEV1^d ≥ 100	1.00 (Reference)		1.00 (Reference)		1.00 (Reference)
ppFEV1^d 80–99	1.94 (1.11–3.38)		1.93 (1.10–3.38)		1.60 (0.90–2.82)
ppFEV1^d 50–79	3.20 (1.79–5.71)		3.09 (1.67–5.71)		2.57 (1.42–4.65)
ppFEV1^d <50	4.10 (1.91–8.78)	<0.001	4.15 (1.89–9.10)	<0.001	3.25 (1.46–7.22)	<0.001
Normal airflow^e	1.00 (Reference)		1.00 (Reference)		1.00 (Reference)
COPD grade 1^f	1.25 (0.67–2.34)		1.20 (0.63–2.27)		1.42 (0.74–2.74)
COPD grade 2^g	1.79 (1.13–2.85)		1.64 (1.00–2.67)		1.68 (1.06–2.67)
COPD grade 3 or 4^h	2.97 (1.59–5.54)	0.001	2.78 (1.46–5.29)	0.007	2.74 (1.45–5.17)	0.002
ppFEV1^d 10% decrease	1.27 (1.17–1.38)		1.27 (1.16–1.38)		1.24 (1.13–1.36)

^a Model 1: Adjusted for age (<40, 40–49,  .  .  .  , ≥ 80 years).

^b Model 2: Adjusted for all in Model 1 + smoking (never, former, current, unknown), and education (<10, 10–12, ≥13 years, unknown).

^c Model 4: Adjusted for all in Model 2 + body mass index [BMI] (<18.5, 18.5–24.9, 25.0–29.9, ≥ 30.0 kg/m²), physical activity (inactive, light activity <1 hours/week, light activity 1–2 hours/week, light activity ≥ 3 hours/week, only vigorous activity, unknown), cardiovascular disease [CVD] (yes, no), diabetes mellitus [DM] (no, yes), systolic blood pressure [SBP] (sex specific quartiles), and total cholesterol (sex specific quartiles).

^d ppFEV₁ denotes% predicted forced expiratory volume in 1 second.

^e Normal airflow defined as forced expiratory volume in 1 second (FEV₁)/forced vital capacity (FVC) ≥70% and% predicted FVC ≥80. Participants with restrictive lung function impairment (FEV₁/FVC ≥ 70% and ppFVC < 80) were excluded from the COPD analyses.

^f COPD grade 1 denotes chronic obstructive pulmonary disease defined as FEV₁/FVC <70% and ppFEV₁ ≥80.

^g COPD grade 2 defined as FEV₁/FVC <70% and 50 ≤ ppFEV₁ < 80.

^h COPD grade 3 or 4 defined as FEV₁/FVC <70% and ppFEV₁ < 50.

eTable 6. Adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) for the associations between respiratory symptoms and all-cause mortality (weighted)

Table

	Model 1:Age	Model 2:Model 1 + smoking, education, BMI, physical activity, CVD	Model 3:Model 2 + lung function	Model 4:Model 3 + DM, SBP, total cholesterol
Respiratory symptom	HR^a (95% CI)^a	HR^b (95% CI)^b	HR^c (95% CI)^c	HR^d (95% CI)^d
Women
Chronic bronchitis	1.70 (1.04–2.79)	1.52 (1.03–2.23)	1.27 (0.85–1.89)	1.27 (0.85–1.89)
Wheeze	1.42 (1.09–1.85)	1.21 (0.91–1.60)	1.05 (0.79–1.40)	1.04 (0.78–1.38)
Dyspnoea uphill	1.52 (0.98–2.35)	1.48 (0.94–2.34)	1.34 (0.83–2.15)	1.25 (0.80–1.97)
Dyspnoea stairs	1.17 (0.76–1.81)	1.07 (0.66–1.74)	0.97 (0.59–1.59)	0.93 (0.56–1.55)
Dyspnoea walking	2.62 (1.69–4.07)	2.43 (1.51–3.92)	1.73 (1.04–2.89)	1.57 (0.92–2.65)
Dyspnoea sitting	3.06 (1.56–5.99)	2.51 (1.26–5.01)	1.94 (0.92–4.09)	1.86 (0.86–4.02)
1 symptom^e	1.54 (1.07–2.22)	1.38 (0.96–2.00)	1.29 (0.90–1.86)	1.24 (0.85–1.80)
2 symptoms^e	2.17 (1.52–3.09)	1.95 (1.38–2.76)	1.36 (0.89–2.07)	1.30 (0.85–1.99)
3 symptoms^e	2.61 (1.75–3.91)	1.75 (1.08–2.84)	1.16 (0.72–1.86)	1.10 (0.67–1.81)
Men
Chronic bronchitis	1.63 (1.18–2.26)	1.49 (1.10–2.02)	1.16 (0.88–1.55)	1.20 (0.91–1.59)
Wheeze	1.32 (1.04–1.68)	1.20 (0.93–1.55)	0.97 (0.74–1.26)	0.97 (0.75–1.27)
Dyspnoea uphill	1.32 (0.93–1.89)	1.27 (0.88–1.84)	1.02 (0.70–1.47)	0.93 (0.64–1.37)
Dyspnoea stairs	1.12 (0.72–1.72)	0.91 (0.56–1.46)	0.76 (0.48–1.22)	0.70 (0.44–1.11)
Dyspnoea walking	2.66 (1.78–3.97)	2.41 (1.59–3.65)	1.57 (1.04–2.36)	1.55 (1.03–2.35)
Dyspnoea sitting	1.62 (0.89–2.96)	1.51 (0.83–2.76)	1.22 (0.68–2.17)	1.28 (0.73–2.26)
1 symptom^e	1.21 (0.87–1.69)	1.14 (0.81–1.62)	0.98 (0.70–1.38)	1.00 (0.71–1.41)
2 symptoms^e	2.03 (1.49–2.74)	1.82 (1.29–2.57)	1.25 (0.88–1.79)	1.31 (0.91–1.88)
3 symptoms^e	3.17 (2.07–4.85)	2.59 (1.64–4.09)	1.48 (1.13–1.36)	1.56 (0.99–2.45)

^a Model 1: Adjusted for age (<40, 40–49,  .  .  .  , ≥80 years).

^b Model 2: Adjusted for all in Model 1 + smoking (never, former, current, unknown), education (<10, 10–12, ≥13 years, unknown), body mass index [BMI] (<18.5, 18.5–24.9, 25.0–29.9, ≥30.0 kg/m²), physical activity (inactive, light activity <1 hours/week, light activity 1–2 hours/week, light activity ≥3 hours/week, only vigorous activity, unknown), and cardiovascular disease [CVD] (yes, no).

^c Model 3: Adjusted for all in Model 2 +% predicted forced expiratory volume in one second (≥100, 80–99, 50–79, <50).

^d Model 4: Adjusted for all in Model 3 + diabetes mellitus [DM] (no, yes), systolic blood pressure [SBP] (sex specific quartiles), and total cholesterol (sex specific quartiles).

^e Symptoms include chronic bronchitis, wheeze, and dyspnoea when walking or sitting.

eTable 7. Sensitivity analyses –participants with CVD at baseline excluded: Adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) for the associations between respiratory symptoms and all-cause mortality (weighted)

Table

	Model 1:Age	Model 2:Model 1 + smoking, education, BMI, physical activity	Model 3:Model 2 + lung function	Model 4:Model 3 + DM, SBP, total cholesterol
Respiratory symptom	HR^a (95% CI)^a	HR^b (95% CI)^b	HR^c (95% CI)^c	HR^d (95% CI)^d
Women
Chronic bronchitis	1.65 (0.94–2.89)	1.42 (0.92–2.19)	1.12 (0.71–1.76)	1.12 (0.71–1.76)
Wheeze	1.45 (1.07–1.96)	1.18 (0.86–1.63)	1.04 (0.75–1.43)	1.05 (0.76–1.44)
Dyspnoea uphill	1.59 (1.00–2.56)	1.51 (0.93–2.47)	1.32 (0.79–2.21)	1.16 (0.70–1.91)
Dyspnoea stairs	1.10 (0.67–1.82)	1.05 (0.62–1.80)	0.89 (0.51–1.57)	0.94 (0.53–1.65)
Dyspnoea walking	2.60 (1.51–4.46)	2.38 (1.39–4.09)	1.54 (0.86–2.76)	1.34 (0.73–2.47)
Dyspnoea sitting	3.03 (1.35–6.76)	2.46 (1.09–5.56)	1.83 (0.75–4.48)	1.75 (0.69–4.43)
1 symptom^e	1.60 (1.07–2.39)	1.38 (0.92–2.07)	1.28 (0.86–1.91)	1.25 (0.83–1.87)
2 symptoms^e	2.26 (1.47–3.48)	1.87 (1.26–2.77)	1.19 (0.70–2.02)	1.13 (0.66–1.93)
3 symptoms^e	2.78 (1.73–4.45)	1.77 (1.00–3.12)	1.06 (0.60–1.89)	1.08 (0.60–1.93)
Men
Chronic bronchitis	1.69 (1.10–2.60)	1.56 (1.04–2.34)	1.19 (0.82–1.72)	1.24 (0.87–1.76)
Wheeze	1.39 (1.05–1.83)	1.26 (0.93–1.71)	1.00 (0.73–1.38)	1.02 (0.74–1.40)
Dyspnoea uphill	1.42 (0.93–2.16)	1.46 (0.98–2.18)	1.22 (0.83–1.80)	1.09 (0.72–1.66)
Dyspnoea stairs	0.88 (0.51–1.51)	0.75 (0.39–1.43)	0.65 (0.35–1.21)	0.59 (0.32–1.08)
Dyspnoea walking	2.31 (1.42–3.75)	2.41 (1.46–3.96)	1.49 (0.96–2.30)	1.47 (0.96–2.23)
Dyspnoea sitting	2.16 (1.64–2.86)	2.02 (1.30–3.14)	1.49 (0.89–2.51)	1.54 (0.87–2.71)
1 symptom^e	1.18 (0.83–1.69)	1.08 (0.73–1.60)	0.95 (0.65–1.39)	0.99 (0.67–1.46)
2 symptoms^e	2.08 (1.40–3.11)	2.04 (1.33–3.14)	1.40 (0.88–2.24)	1.41 (0.89–2.23)
3 symptoms^e	3.49 (1.97–6.19)	2.90 (1.67–5.05)	1.47 (0.83–2.61)	1.55 (0.87–2.75)

^a Model 1: Adjusted for age (<40, 40–49,  .  .  .  , ≥80 years).

^b Model 2: Adjusted for all in Model 1 + smoking (never, former, current, unknown), education (<10, 10–12, ≥13 years, unknown), body mass index [BMI] (<18.5, 18.5–24.9, 25.0–29.9, ≥30.0 kg/m²), and physical activity (inactive, light activity <1 hours/week, light activity 1–2 hours/week, light activity ≥3 hours/week, only vigorous activity, unknown).

^c Model 3: Adjusted for all in Model 2 +% predicted forced expiratory volume in 1 second (≥100, 80–99, 50–79, <50).

^d Model 4: Adjusted for all in Model 3 + diabetes mellitus [DM] (no, yes), systolic blood pressure [SBP] (sex specific quartiles), and total cholesterol (sex specific quartiles).

^e Symptoms include chronic bronchitis, wheeze, and dyspnoea when walking or sitting.

eTable 8. Death rates and adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) for the associations between respiratory symptoms and cardiovascular mortality (weighted) in participants without CVD at baseline

Table

					Model 1:Age	Model 2:Model 1 + smoking, education, BMI, physical activity	Model 3:Model 2 + lung function
Respiratory symptom	Person-years	Deaths	Death rate^a	(95% CI)	HR^b	HR^c (95% CI)^c	HR^d (95% CI)^d
Women
No chronic bronchitis	331,428	953	2.88	(2.22–3.78)	1.00	1.00 (Reference)	1.00 (Reference)
Chronic bronchitis	10,477	64	6.11	(2.47–19.21)	1.68	1.65 (0.74–3.68)	1.67 (0.74–3.75)
No wheeze	284,715	830	2.92	(2.19–3.97)	1.00	1.00 (Reference)	1.00 (Reference)
Wheeze	60,573	176	2.91	(1.94–4.54)	1.17	0.90 (0.51–1.57)	0.87 (0.50–1.54)
No dyspnoea	188,298	464	2.46	(1.68–3.76)	1.00	1.00 (Reference)	1.00 (Reference)
Dyspnoea uphill	27,048	121	4.47	(2.20–10.56)	1.58	1.13 (0.49–2.58)	1.18 (0.51–2.77)
Dyspnoea stairs	27,842	114	4.09	(2.06–9.34)	1.14	1.04 (0.43–2.50)	1.06 (0.41–2.72)
Dyspnoea walking	6918	53	7.66	(3.29–21.42)	2.16	1.31 (0.60–2.88)	1.24 (0.55–2.80)
Dyspnoea sitting	3602	7	1.94	(0.90–4.82)	1.31	0.83 (0.36–1.90)	0.87 (0.36–2.14)
No symptoms^e	197,654	562	2.84	(2.00–4.18)	1.00	1.00 (Reference)	1.00 (Reference)
1 symptom	42,401	113	2.67	(1.44–5.48)	1.07	0.71 (0.34–1.51)	0.69 (0.33–1.44)
2 symptoms	7343	47	6.40	(2.48–21.41)	2.77	1.79 (0.70–4.57)	1.99 (0.71–5.57)
3 symptoms	1566	8	5.11	(2.37–12.11)	2.39	1.40 (0.54–3.67)	1.28 (0.46–3.56)
Men
No chronic bronchitis	258,371	1044	4.04	(3.17–5.24)	1.00	1.00 (Reference)	1.00 (Reference)
Chronic bronchitis	10,279	75	7.30	(3.34–18.83)	1.35	1.32 (0.63–2.79)	1.01 (0.47–2.16)
No wheeze	222,425	848	3.81	(2.88–5.16)	1.00	1.00 (Reference)	1.00 (Reference)
Wheeze	49,596	282	5.69	(3.92–8.54)	1.43	1.34 (0.84–2.16)	1.06 (0.63–1.80)
No dyspnoea	167,981	622	3.70	(2.67–5.30)	1.00	1.00 (Reference)	1.00 (Reference)
Dyspnoea uphill	16,191	104	6.42	(3.27–14.24)	1.33	1.50 (0.70–3.18)	1.20 (0.54–2.66)
Dyspnoea stairs	11,989	58	4.84	(2.18–12.52)	0.73	0.77 (0.29–2.03)	0.65 (0.24–1.73)
Dyspnoea walking	3501	62	17.71	(7.19–55.04)	2.44	2.80 (1.09–7.18)	1.90 (0.85–4.26)
Dyspnoea sitting	2542	6	2.36	(1.03–6.39)	0.41	0.41 (0.13–1.29)	0.30 (0.09–0.99)
No symptoms^e	155,958	556	3.57	(2.52–5.22)	1.00	1.00 (Reference)	1.00 (Reference)
1 symptom	33,355	215	6.45	(4.00–11.04)	1.53	1.45 (0.78–2.70)	1.27 (0.69–2.33)
2 symptoms	6226	68	10.92	(4.74–31.05)	2.12	2.09 (0.84–5.16)	1.34 (0.48–3.75)
3 symptoms	980	7	7.15	(3.46–16.98)	1.16	1.12 (0.39–3.23)	0.61 (0.19–1.90)

^a Per 1000 person-years.

^b Model 1: Adjusted for age (<40, 40–49,  .  .  .  , ≥80 years).

^c Model 2: Adjusted for all in Model 1 + smoking (never, former, current, unknown), education (<10, 10–12, ≥13 years, unknown), body mass index [BMI] (<18.5, 18.5–24.9, 25.0–29.9, ≥30.0 kg/m²), and physical activity (inactive, light activity <1 hours/week, light activity 1–2 hours/week, light activity ≥3 hours/week, only vigorous activity, unknown).

^d Model 3: Adjusted for all in Model 2 +% predicted forced expiratory volume in 1 second (≥100, 80–99, 50–79, <50).

^e Symptoms include chronic bronchitis, wheeze, and dyspnoea when walking or sitting.

eTable 9. Adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) for the associations between respiratory symptoms and cardiovascular mortality (weighted) in participants without CVD at baseline

Table

	Model 1:Age	Model 2:Model 1 + smoking, education, BMI, physical activity	Model 3:Model 2 + lung function	Model 4:Model 3 + DM, SBP, total cholesterol
Respiratory symptom	HR^a (95% CI)^a	HR^b (95% CI)^b	HR^c (95% CI)^c	HR^d (95% CI)^d
Women
Chronic bronchitis	1.68 (0.57–4.96)	1.65 (0.74–3.68)	1.67 (0.74–3.75)	1.78 (0.75–4.20)
Wheeze	1.17 (0.69–1.97)	0.90 (0.51–1.57)	0.87 (0.50–1.54)	0.85 (0.47–1.55)
Dyspnoea uphill	1.58 (0.69–3.58)	1.13 (0.49–2.58)	1.18 (0.51–2.77)	1.11 (0.43–2.84)
Dyspnoea stairs	1.14 (0.50–2.64)	1.04 (0.43–2.50)	1.06 (0.41–2.72)	1.22 (0.49–3.05)
Dyspnoea walking	2.16 (1.04–4.47)	1.31 (0.60–2.88)	1.24 (0.55–2.80)	1.17 (0.42–3.23)
Dyspnoea sitting	1.31 (0.60–2.85)	0.83 (0.36–1.90)	0.87 (0.36–2.14)	0.86 (0.29–2.56)
1 symptom^e	1.07 (0.54–2.15)	0.71 (0.34–1.51)	0.69 (0.33–1.44)	0.67 (0.29–1.55)
2 symptoms^e	2.77 (1.14–6.70)	1.79 (0.70–4.57)	1.99 (0.71–5.57)	1.90 (0.67–5.40)
3 symptoms^e	2.39 (1.05–5.61)	1.40 (0.54–3.67)	1.28 (0.46–3.56)	1.51 (0.49–4.60)
Men
Chronic bronchitis	1.35 (0.64–2.82)	1.32 (0.63–2.79)	1.01 (0.47–2.16)	0.97 (0.44–2.15)
Wheeze	1.43 (0.90–2.26)	1.34 (0.84–2.16)	1.06 (0.63–1.80)	1.08 (0.64–1.82)
Dyspnoea uphill	1.33 (0.65–2.73)	1.50 (0.70–3.18)	1.20 (0.54–2.66)	1.21 (0.47–3.11)
Dyspnoea stairs	0.73 (0.30–1.75)	0.77 (0.29–2.03)	0.65 (0.24–1.73)	0.54 (0.23–1.27)
Dyspnoea walking	2.44 (1.13–5.28)	2.80 (1.09–7.18)	1.90 (0.85–4.26)	2.03 (0.87–4.77)
Dyspnoea sitting	0.41 (0.14–1.20)	0.41 (0.13–1.29)	0.30 (0.09–0.99)	0.30 (0.08–1.04)
1 symptom^e	1.53 (0.85–2.73)	1.45 (0.78–2.70)	1.27 (0.69–2.33)	1.29 (0.69–2.41)
2 symptoms^e	2.12 (0.93–4.86)	2.09 (0.84–5.16)	1.34 (0.48–3.75)	1.25 (0.41–3.82)
3 symptoms^e	1.16 (0.45–2.98)	1.12 (0.39–3.23)	0.61 (0.19–1.90)	0.63 (0.21–1.90)

^a Model 1: Adjusted for age (<40, 40–49,  .  .  .  , ≥80 years).

^b Model 2: Adjusted for all in Model 1 + smoking (never, former, current, unknown), education (<10, 10–12, ≥13 years, unknown), body mass index [BMI] (<18.5, 18.5–24.9, 25.0–29.9, ≥30.0 kg/m²), and physical activity (inactive, light activity <1 hours/week, light activity 1–2 hours/week, light activity ≥3 hours/week, only vigorous activity, unknown).

^c Model 3: Adjusted for all in Model 2 +% predicted forced expiratory volume in one second (≥100, 80–99, 50–79, <50).

^d Model 4: Adjusted for all in Model 3 + diabetes mellitus [DM] (no, yes), systolic blood pressure [SBP] (sex specific quartiles), and total cholesterol (sex specific quartiles).

^e Symptoms include chronic bronchitis, wheeze, and dyspnoea when walking or sitting.

eTable 10. Sensitivity analyses –participants with CVD at baseline included: Adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) for the associations between respiratory symptoms and cardiovascular mortality (weighted)

Table

	Model 1:Age	Model 2:Model 1 + smoking, education, BMI, physical activity, CVD	Model 3:Model 2 + lung function	Model 4:Model 3 + DM, SBP, total cholesterol
Respiratory symptom	HR^a (95% CI)^a	HR^b (95% CI)^b	HR^c (95% CI)^c	HR^d (95% CI)^d
Women
Chronic bronchitis	1.50 (0.63–3.55)	1.36 (0.67–2.75)	1.33 (0.66–2.70)	1.34 (0.62–2.89)
Wheeze	1.04 (0.68–1.59)	0.78 (0.49–1.25)	0.75 (0.46–1.21)	0.74 (0.45–1.21)
Dyspnoea uphill	1.46 (0.70–3.05)	1.20 (0.57–2.49)	1.23 (0.58–2.62)	1.20 (0.55–2.61)
Dyspnoea stairs	1.33 (0.69–2.58)	0.99 (0.46–2.12)	1.00 (0.46–2.21)	1.01 (0.45–2.26)
Dyspnoea walking	2.71 (1.49–4.93)	1.74 (0.95–3.20)	1.75 (0.93–3.32)	1.67 (0.86–3.26)
Dyspnoea sitting	1.58 (0.85–2.97)	0.95 (0.45–2.02)	0.98 (0.45–2.14)	0.83 (0.28–2.47)
1 symptom^e	1.11 (0.62–1.97)	0.82 (0.45–1.49)	0.81 (0.45–1.48)	0.80 (0.43–1.50)
2 symptoms^e	2.40 (1.27–4.54)	1.65 (0.81–3.34)	1.71 (0.81–3.64)	1.77 (0.83–3.76)
3 symptoms^e	2.27 (1.24–4.17)	1.01 (0.46–2.21)	0.99 (0.45–2.18)	0.83 (0.27–2.53)
Men
Chronic bronchitis	1.33 (0.81–2.18)	1.14 (0.71–1.83)	0.89 (0.54–1.45)	0.88 (0.55–1.42)
Wheeze	1.38 (0.96–1.98)	1.24 (0.85–1.81)	0.99 (0.67–1.47)	0.96 (0.65–1.42)
Dyspnoea uphill	1.27 (0.71–2.28)	1.03 (0.56–1.91)	0.80 (0.41–1.55)	0.73 (0.36–1.51)
Dyspnoea stairs	0.93 (0.48–1.80)	0.71 (0.36–1.38)	0.57 (0.29–1.09)	0.46 (0.25–0.87)
Dyspnoea walking	2.78 (1.58–4.88)	2.14 (1.10–4.17)	1.48 (0.79–2.79)	1.44 (0.74–2.80)
Dyspnoea sitting	0.59 (0.25–1.38)	0.36 (0.11–1.13)	0.31 (0.11–0.87)	0.34 (0.12–0.91)
1 symptom^e	1.48 (0.91–2.39)	1.33 (0.79–2.22)	1.13 (0.68–1.88)	1.13 (0.67–1.90)
2 symptoms^e	2.26 (1.27–4.03)	1.84 (0.96–3.54)	1.27 (0.63–2.55)	1.21 (0.61–2.43)
3 symptoms^e	1.53 (0.77–3.04)	1.19 (0.58–2.44)	0.77 (0.35–1.72)	0.84 (0.38–1.84)

^a Model 1: Adjusted for age (<40, 40–49,  .  .  .  , ≥80 years).

^b Model 2: Adjusted for all in Model 1 + smoking (never, former, current, unknown), education (<10, 10–12, ≥13 years, unknown), body mass index [BMI] (<18.5, 18.5–24.9, 25.0–29.9, ≥30.0 kg/m²), physical activity (inactive, light activity <1 hours/week, light activity 1–2 hours/week, light activity ≥3 hours/week, only vigorous activity, unknown), and cardiovascular disease (no, yes).

^c Model 3: Adjusted for all in Model 2 +% predicted forced expiratory volume in 1 second (≥100, 80–99, 50–79, <50).

^d Model 4: Adjusted for all in Model 3 + diabetes mellitus [DM] (no, yes), systolic blood pressure [SBP] (sex specific quartiles), and total cholesterol (sex specific quartiles).

^e Symptoms include chronic bronchitis, wheeze, and dyspnoea when walking or sitting.

eTable 11. Additional analyses –women and men combined: Death rates and adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) for the associations between dyspnoea and cardiovascular mortality (weighted)

Table

				Model 1:Age, sex	Model 2:Model 1 + smoking, education, BMI, physical activity	Model 3:Model 2 + lung function	Model 4:Model 3 + DM, SBP, total cholesterol
Respiratory symptom	Person-years	Deaths	Death rate^a (95% CI)	HR^b (95% CI)^c	HR^c (95% CI)^c	HR^d (95% CI)^d	HR^e (95% CI)^e
Excluding participants with CVD at baseline
No dyspnoea	356,279	1086	3.05 (2.37–3.98)	1.00 (Reference)	1.00 (Reference)	1.00 (Reference)	1.00 (Reference)
Dyspnoea uphill	43,239	225	5.20 (3.21–9.01)	1.43 (0.83–2.48)	1.26 (0.72–2.18)	1.15 (0.66–2.03)	1.05 (0.57–1.94)
Dyspnoea stairs	3,830	172	4.32 (2.58–7.76)	0.94 (0.51–1.73)	0.90 (0.46–1.76)	0.84 (0.41–1.71)	0.87 (0.43–1.76)
Dyspnoea walking	10,418	115	11.04 (6.10–21.93)	2.27 (1.33–3.86)	1.86 (0.98–3.55)	1.48 (0.78–2.79)	1.54 (0.78–3.02)
Dyspnoea sitting	6144	13	2.12 (1.21–3.98)	0.67 (0.32–1.41)	0.52 (0.20–1.32)	0.42 (0.16–1.13)	0.42 (0.15–1.12)
Including participants with CVD at baseline
No dyspnoea	368,252	1415	3.84 (3.09–4.85)	1.00 (Reference)	1.00 (Reference)	1.00 (Reference)	1.00 (Reference)
Dyspnoea uphill	48,290	316	6.54 (4.38–10.21)	1.35 (0.85–2.15)	1.11 (0.69–1.78)	1.00 (0.61–1.63)	0.93 (0.56–1.55)
Dyspnoea stairs	45,797	324	7.07 (4.85–10.72)	1.13 (0.71–1.80)	0.85 (0.51–1.43)	0.80 (0.47–1.36)	0.73 (0.42–1.27)
Dyspnoea walking	12,754	239	18.74 (12.65–28.80)	2.74 (1.82–4.14)	2.03 (1.28–3.21)	1.67 (1.05–2.66)	1.68 (1.03–2.73)
Dyspnoea sitting	6977	29	4.16 (2.78–6.37)	0.80 (0.41–1.56)	0.46 (0.18–1.16)	0.43 (0.18–1.01)	0.42 (0.18–1.00)

^a Per 1000 person-years.

^b Model 1: Adjusted for age (<40, 40–49,  .  .  .  , ≥80 years) and sex (women, man).

^c Model 2: Adjusted for all in Model 1 + smoking (never, former, current, unknown), education (<10, 10–12, ≥13 years, unknown), body mass index [BMI] (<18.5, 18.5–24.9, 25.0–29.9, ≥30.0 kg/m²), and physical activity (inactive, light activity <1 hours/week, light activity 1–2 hours/week, light activity ≥3 hours/week, only vigorous activity, unknown).

^d Model 3: Adjusted for all in Model 2 +% predicted forced expiratory volume in one second (≥100, 80–99, 50–79, <50).

^e Model 4: Adjusted for all in Model 3 + diabetes mellitus [DM] (no, yes), systolic blood pressure [SBP] (sex specific quartiles), and total cholesterol (sex specific quartiles).

eTable 12. Respiratory symptoms, lung function, medical diagnoses, and other characteristics of participants with and without data on the dyspnoea scale

Table

	Women		Men
	Data on dyspnoea scale73.0%^a	Missing data on dyspnoea scale27.0%^a	Data on dyspnoea scale73.8%^a	Missing data on dyspnoea scale26.2%^a
Respiratory symptoms	%^b (95% CI)	%^b (95% CI)	%^b (95% CI)	%^b (95% CI)
Chronic bronchitis	3.5 (2.8–4.3)	3.1 (1.8–4.4)	4.7 (3.7–5.7)	3.1 (1.9–4.2)
Wheeze	19.1 (17.3–21.0)	16.8 (14.0–19.6)	19.9 (17.9–21.9)	18.0 (14.8–21.2)
Lung function	Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)
FEV1^c [L]	2.8 (0.7)	2.8 (0.7)	3.8 (1.0)	3.8 (1.0)
ppFEV1^d	96.6 (16.3)	96.9 (15.6)	94.7 (16.8)	93.8 (15.5)
FVC^e	3.5 (0.8)	3.5 (0.8)	4.9 (1.0)	4.9 (1.0)
FEV1/FVC [%]	79.9 (7.4)	80.0 (7.3)	76.8 (8.9)	77.0 (8.0)
Lung function	%^b (95% CI)	%^b (95% CI)	%^b (95% CI)	%^b (95% CI)
ppFEV1^d ≥ 100	43.4 (40.6–46.2)	41.7 (37.0–46.3)	37.9 (34.9–40.8)	34.7 (30.8–39.7)
ppFEV1^d 80–99	43.0 (40.2–45.8)	46.5 (41.8–51.2)	46.8 (43.8–49.8)	50.2 (45.0–55.3)
ppFEV1^d 50–79	12.6 (10.8–14.3)	11.3 (8.5–14.0)	13.6 (11.7–15.6)	13.9 (10.6–17.1)
ppFEV1^d <50	1.0 (0.6–1.4)	0.6 (0.4–0.7)	1.7 (1.2–2.2)	1.2 (0.5–1.9)
Normal airflow ^f	86.5 (84.8–88.3)	86.4 (83.4–89.4)	80.3 (78.0–82.6)	81.9 (78.2–85.6)
COPD grade 1^g	2.4 (1.6–3.3)	3.7 (1.9–5.4)	6.5 (5.0–8.0)	5.1 (2.8–7.3)
COPD grade 2^h	5.5 (4.4–6.6)	5.5 (3.5–7.5)	8.5 (7.0–10.1)	9.3 (6.6–12.1)
COPD grade 3 or 4^i	1.0 (0.6–1.4)	0.6 (0.4–0.7)	1.6 (0.1–2.0)	0.9 (0.6–1.1)
Medical diagnoses (ever had)^b	%^b (95% CI)	%^b (95% CI)	%^b (95% CI)	%^b (95% CI)
Myocardial infarction	1.1 (0.6–1.7)	1.1 (0.2–2.0)	4.9 (3.6–6.1)	6.5 (4.0–8.9)
Angina pectoris	2.2 (1.5–2.9)	3.3 (1.7–4.9)	4.7 (3.5–5.9)	6.7 (4.2–9.1)
Stroke or cerebral haemorrhage	1.7 (1.0–2.4)	1.6 (0.5–2.8)	2.1 (1.2–2.9)	1.1 (0.2–2.0)
Diabetes mellitus	2.5 (1.6–3.3)	2.3 (1.0–3.7)	2.2 (1.4–3.1)	2.9 (1.3–4.6)
Other	Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)
Age [years]	49.9 (15.9)	49.6 (17.2)	51.2 (15.8)	49.0 (16.5)
Body mass index [kg/m^2]	26.2 (4.4)	26.3 (4.6)	26.4 (3.3)	26.6 (3.6)
Systolic blood pressure [mmHg]	135.6 (23.3)	135.7 (25.3)	140.6 (19.0)	140.4 (20.1)
Total cholesterol [mmol/l]	6.0 (1.3)	6.0 (1.3)	5.9 (1.1)	5.8 (1.1)
Other^b	%^b (95% CI)	%^b (95% CI)	%^b (95% CI)	%^b (95% CI)
Education ≥13 years	19.5 (17.2–21.9)	21.4 (17.4–25.4)	19.4 (16.9–21.9)	18.3 (14.2–22.5)
Ever smoker	51.1 (48.2–54.0)	47.9 (43.1–52.8)	61.3 (58.3–64.4)	62.8 (57.7–68.0)
Inactive	6.7 (5.2–8.1)	7.2 (4.7–9.7)	6.6 (5.1–8.1)	9.2 (6.1–12.2)

^a Percent within sex.

^b Percent within variable and sex, missing excluded.

^c FEV₁ denotes forced expiratory volume in 1 second.

^d ppFEV₁ denotes% predicted FEV₁.

^e FVC denotes forced vital capacity.

^f Normal airflow defined by FEV₁/FVC ≥ 70% and% predicted FVC ≥ 80.

^g COPD Grade 1 denotes chronic obstructive pulmonary disease defined as FEV₁/FVC <70% and ppFEV₁ ≥ 80.

^h COPD Grade 2 defined as FEV₁/FVC <70% and 80 > ppFEV₁ ≥ 50.

ⁱ COPD Grade 3 or 4 defined as FEV₁/FVC <70% and ppFEV₁ <50.

eTable 13. Adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) for the associations of having missing data on the dyspnoea scale with all-cause and cardiovascular mortality (weighted)

	All-cause mortality		Cardiovascular mortality^a
	Women	Men	Women	Men
Models^b	HR (95% CI)	HR (95% CI)	HR (95% CI)	HR (95% CI)
Model 1: Age	1.15 (0.87–1.53)	1.09 (0.83–1.44)	0.88 (0.50–1.56)	1.18 (0.70–2.01)
Model 2: Model 1 + smoking, education, BMI, physical activity, CVD	1.09 (0.81–1.45)	1.05 (0.79–1.40)	0.85 (0.46–1.55)	1.22 (0.70–2.12)
Model 3: Model 2 + lung function	1.11 (0.83–1.47)	1.10 (0.83–1.46)	0.87 (0.48–1.58)	1.38 (0.80–2.36)
Model 4: Model 3 + DM, SBP, total cholesterol	1.13 (0.85–1.51)	1.06 (0.80–1.41)	0.91 (0.50–1.66)	1.37 (0.77–2.42)

^a Participants with baseline cardiovascular disease excluded.

^b Categorization of possible confounders: Age (<40, 40–49,  .  .  .  , ≥80 years), smoking (never, former, current, unknown), education (<10, 10–12, ≥13 years, unknown), body mass index [BMI] (<18.5, 18.5–24.9, 25.0–29.9, ≥30.0 kg/m²), physical activity (inactive, light activity <1 hours/week, light activity 1–2 hours/week, light activity ≥3 hours/week, only vigorous activity, unknown), cardiovascular disease [CVD] (yes, no),% predicted forced expiratory volume in 1 second (≥100, 80–99, 50–79, <50), diabetes mellitus [DM] (no, yes), systolic blood pressure [SBP] (sex specific quartiles), and total cholesterol (sex specific quartiles).

Figure 1a.  Association of the combined exposure of% predicted forced expiratory volume in 1 second (ppFEV₁) and reporting of chronic bronchitis with all-cause mortality. The reference category consists of participants with ppFEV₁ ≥ 100 not reporting chronic bronchitis. The bars represent 95% confidence intervals. Adjusted for age (<40, 40–49,  .  .  .  , ≥ 80 years), sex (woman, man), smoking (never, former, current, unknown), education (<10, 10–12, ≥13 years, unknown), body mass index (<18.5, 18.5–24.9, 25.0–29.9, ≥ 30.0 kg/m²), physical activity (inactive, light activity <1 hours/week, light activity 1–2 hours/week, light activity ≥3 hours/week, only vigorous activity, unknown), and cardiovascular disease (yes, no).

Figure 1b.  Association of the combined exposure of% predicted forced expiratory volume in 1 second (ppFEV₁) and reporting of wheeze with all-cause mortality. The reference category consists of participants with ppFEV₁≥ 100 not reporting wheeze. The bars represent 95% confidence intervals. Adjusted for age (<40, 40–49,  .  .  .  , ≥80 years), sex (woman, man), smoking (never, former, current, unknown), education (<10, 10–12, ≥13 years, unknown), body mass index (<18.5, 18.5–24.9, 25.0–29.9, ≥30.0 kg/m²), physical activity (inactive, light activity <1 hours/week, light activity 1–2 hours/week, light activity ≥ 3 hours/week, only vigorous activity, unknown), and cardiovascular disease (yes, no).