This study corresponds to a sub-study of the ILERVAS project (NCT03228459) [22], a prospective, multicenter, longitudinal cohort conducted in primary care centers across the province of Lleida (Catalonia, Spain) between 2015 and 2018. The overarching objective of the ILERVAS project was to characterize the prevalence of subclinical atheromatous disease and previously undiagnosed chronic kidney disease in a population with low to moderate cardiovascular risk.

The primary objective of our study was to prospectively evaluate the relationship between changes in lung function, categorized by tertiles, and the progression of subclinical atherosclerosis over a 4-year follow-up period. The secondary objective was to determine whether the development of AO during the 4-year follow-up period was specifically associated with the progression of subclinical atherosclerosis.

This was an observational longitudinal study using data from the ILERVAS cohort [22]. Baseline assessments were conducted between 2015 and 2018, and follow-up visits took place approximately 4 years later, between 2019 and 2021, with a mean follow-up duration of 3.99 (SD, 0.22) years. The inclusion criteria for this sub-study were: (1) women aged 50–70 years and men aged 45–65 years; and (2) presence of at least one cardiovascular risk factor (such as hypertension, dyslipidemia, obesity, smoking history, or a first-degree family history of premature cardiovascular disease), with no previous clinical history of cardiovascular disease, diabetes, chronic kidney disease, active neoplasia, or a life expectancy of less than 18 months.

Among 8330 individuals initially enrolled in the ILERVAS cohort, 1743 had both baseline and follow-up spirometry available for longitudinal analyses. Extreme pulmonary function values were identified using symmetric trimming at the ≤0.5th and ≥99.5th percentiles. All spirometric values meeting these criteria were individually reviewed by a pulmonologist, who confirmed that they were not physiologically plausible and were most consistent with technical artifacts or nonreproducible maneuvers. Fifty participants were excluded at this stage, resulting in 1693 individuals.

Subsequently, extreme values in annual change of total atheromatous plaque area were identified using the same symmetric trimming approach (≤0.5th and ≥99.5th percentiles). Seventy additional participants were excluded based on this criterion, yielding a final analytic sample of 1623 participants (Fig. 1).

Fig. 1.

Flowchart of longitudinal cohort and incident airflow obstruction analyses.

Within this final analytic sample, participants with AO at baseline were excluded from incident analyses. According to the primary fixed-ratio definition (FEV1/FVC<0.70), 194 individuals (12%) met criteria for baseline obstruction and were therefore not eligible for incident evaluation. The remaining 1429 participants were eligible for incident analysis. During follow-up, 160 developed incident AO, whereas 1269 remained nonobstructed.

In predefined sensitivity analyses using the lower limit of normal (LLN) definition (FEV1/FVC z-score<−1.645), incident AO was assessed among participants without baseline LLN-defined obstruction. Under this alternative definition, 1517 individuals were eligible for incident evaluation, of whom 116 (7.65%) developed AO during follow-up.

The ILERVAS study protocol was approved by the Ethics Committee of the Hospital Universitario Arnau de Vilanova (CEIC-1410; CEIC-2015), and all participants provided written informed consent. The study was conducted in accordance with the Declaration of Helsinki.

No formal sample size calculation was performed for this ancillary analysis; all eligible participants from the ILERVAS cohort meeting the inclusion criteria were included.

In the current study, sociodemographic and anthropometric variables were collected, including sex, age, and body mass index (BMI, in kg/m2). Comorbidities such as hypertension, dyslipidemia, obesity, and COPD (based on self-report of a prior clinical diagnosis) were documented, as well as lifestyle factors, including tobacco use. In addition, capillary blood parameters and pharmacotherapy data were collected.

Forced spirometry was performed using a portable ultrasonic spirometer (Datospir©, Sibelmed, Barcelona, Spain) by certified pulmonary technicians, following the standards established by the European Respiratory Society (ERS) and the Spanish Society of Pulmonology and Thoracic Surgery (SEPAR) [23,24]. Each participant completed at least three acceptable and reproducible maneuvers, and the highest values of forced vital capacity (FVC) and forced expiratory volume in the first second (FEV1) were selected for analysis. Postbronchodilator testing was not performed. FVC and FEV1 were recorded in absolute values (mL) and expressed as a percentage of predicted values based on ERS reference equations, as well as z-scores derived from Global Lung Function Initiative (GLI) reference equations.

AO was defined primarily using a fixed FEV1/FVC ratio<0.70, in accordance with the Global Initiative for Chronic Obstructive Lung Disease (GOLD) criteria [16]. For sensitivity analyses, AO was alternatively defined using the LLN, corresponding to an FEV1/FVC z-score<−1.645 derived from GLI reference equations [25].

Pulmonary function trajectories were examined first by tertiles of annual FVC and FEV1 change, and subsequently according to the FEV1/FVC ratio at baseline and at follow-up. Two groups were defined: (1) incident AO group: participants with FEV1/FVC ≥0.70 at baseline who declined to <0.70 at follow-up; and (2) nonincident AO group: participants who maintained FEV1/FVC≥0.70 at both time points.

Arterial ultrasound was performed in 12 vascular territories: both carotid arteries (common, bifurcation, internal, and external) and femoral arteries (common and superficial) [26]. Trained sonographers acquired the images using a VIVID-i ultrasound system (BT09 model, GE Healthcare, Waukesha, WI, USA), equipped with a 12L-RS linear broadband probe (6–13MHz) and pulsed Doppler capabilities to assess hemodynamic abnormalities. Standardized and validated scanning and reading protocols were followed to minimize interoperator variability and type II errors. All readings were performed by blinded evaluators with no access to participants’ clinical data.

Subclinical atherosclerosis was defined as the presence of at least one plaque in any of the 12 assessed regions. According to the Mannheim consensus, a plaque was defined as a focal thickening of the intima-media complex protruding into the lumen ≥1.5mm [27]. The total plaque area (mm2) was measured for all identified plaques [20,28].

For the purpose of the study, the primary outcome was the annual change in total plaque area (carotid and femoral territories).

At the 4-year follow-up visit, participants underwent repeated evaluations of sociodemographic and anthropometric characteristics, pulmonary function (spirometry), and subclinical atheromatous plaque burden through vascular ultrasound.

Baseline characteristics were summarized for all individuals and separately by study groups (incident AO group and nonincident AO group). Continuous variables were described as means (SD,) or medians (IQR,) according to their distribution, while categorical variables were expressed as frequencies (percentages). Group comparisons were performed using the Student t test or the Mann–Whitney U test for continuous variables, as appropriate. The Mann–Whitney U test evaluates differences in the overall distribution of values rather than differences in medians alone. Categorical variables were compared using χ2 tests. Pulmonary function and atheromatous plaque data (affected vascular territories and plaque area) were also summarized at baseline and as annual change for all individuals and by study group using the same descriptive approach.

To evaluate the association between annual change in pulmonary function or incident airflow obstruction and annual change in total atheromatous plaque area (total, carotid, and femoral territories), multivariable linear regression models were fitted. Although baseline plaque area measures are right-skewed, annual change in total atheromatous plaque area is a continuous variable that can assume both positive and negative values, supporting the use of linear regression to obtain clinically interpretable effect estimates.

All models were adjusted for potential confounders, including age, sex, systolic and diastolic blood pressure, total cholesterol, body mass index (BMI), waist circumference, smoking status, pack-years of smoking, antihypertensive treatment, statin therapy, glycated hemoglobin (HbA1c), and baseline total plaque area. Medication variables were defined at baseline.

Analyses were additionally stratified by smoking status, defined as non-smokers (never and former smokers combined) versus current smokers.

A sensitivity analysis was performed by rerunning the fully adjusted main models without percentile trimming of extreme annual plaque-change values, using the same model specifications, to assess the robustness of the findings to the inclusion of extreme values.

Further sensitivity analyses were conducted using the LLN-based definition of incident AO to explore its association with annual change in total atheromatous plaque area (total, carotid, and femoral territories), applying the same multivariable linear regression approach and adjustments.

Multicollinearity was assessed using variance inflation factors (VIF). Model assumptions were evaluated through residual diagnostics, with particular attention to heteroscedasticity. All analyses were conducted using heteroscedasticity-robust (Huber–White) standard errors. Adjusted estimates with 95%CI were reported.

All analyses were performed using R (version 4.4; R Foundation for Statistical Computing, Vienna, Austria). A 2-sided P<.05 was considered statistically significant.

Baseline characteristics of the cohort (N=1623) are summarized in Table 1. Women represented 52.5% of participants, with a mean age of 57.4 (SD, 6.27) years. The most prevalent comorbidities were dyslipidemia (51.6%) and hypertension (39.2%); 25.5% were current smokers and 39.9% were never smokers.

Median annual changes in pulmonary function were −95.39mL (IQR, −172.19 to −24.76) and −2.04% (IQR, −4.19 to 0.24) for FVC, and −77.82mL (IQR, −136.43 to −22.60) and −2.00% (IQR, −4.20 to 0.13) for FEV1 (e-Table 1, in supplementary material). Regarding atheromatous plaque burden, the annual median change was 0.25 (IQR, 0.00–0.50) affected territories, 6.20mm2 (IQR, 1.00–15.3) for total plaque area, and 1.44mm2 (IQR, 0.00–10.8) for carotid plaque area (e-Table 2, in supplementary material).

Groups defined by tertiles of annual FVC and FEV1 change were generally comparable at baseline (e-Table 3, in supplementary material). A modest difference was noted in the proportion of female participants across FEV1 tertiles (P=.041). Across FVC tertiles, greater declines in pulmonary function were significantly associated with a higher atherosclerotic burden, with the third tertile showing more territories with plaque (2.0 [0.0–3.0] vs 1.0 [0.0–3.0]; P=.019) and larger carotid plaque area (25.0mm2 [0.0–72.5] vs 17.0mm2 [0.0–60.0]; P=.018). For FEV1, differences were less marked but suggested greater plaque involvement and larger carotid burden with steeper declines.

After adjustment for potential confounders (sex, age, systolic blood pressure, diastolic blood pressure, total cholesterol, BMI, waist circumference, smoking status, pack-years of smoking, antihypertensive treatment, statin therapy, HbA1c, and baseline plaque burden), annual changes in pulmonary function categorized by tertiles were not significantly associated with plaque progression (e-Fig. 1, in supplementary material).

Of the 1623 participants included in the longitudinal analyses, 1429 had a normal FEV1/FVC ratio at baseline and were eligible for the incident AO analysis. At follow-up, 160 participants (11.2%) developed AO. They were slightly older (58.7 [SD, 6.33] vs 56.9 [SD, 6.18] years; P=.001) and more frequently women (61.9% vs 51.4%; P=.015) compared with those who did not. No significant differences in cardiometabolic comorbidities or smoking status were observed between groups (Table 2).

At baseline, FEV1 (mL) and the FEV1/FVC ratio were lower in the incident AO group, whereas FVC (mL) was similar. Percent-predicted FVC was slightly higher in this group, with no difference in percent-predicted FEV1. Longitudinally, incident AO was associated with a greater annual decline in FEV1 (−146.73 vs −74.90mL; P<.001) and percent-predicted FEV1 (−4.69% vs −1.84%; P<.001). In contrast, annual FVC decline (both absolute and percent-predicted values) was greater in the nonincident AO group (−97.14 vs −72.08mL and −2.07% vs −1.21%, respectively). The FEV1/FVC ratio declined significantly in the incident AO group (−2.50; P<.001) but remained stable in the nonincident AO group (Table 3).

Baseline plaque burden was comparable between both groups. However, over the follow-up period, the incident AO group showed a greater annual increase in plaque extent, with more affected vascular territories (0.26 [IQR, 0.00–0.75] vs 0.25 [IQR, 0.00–0.50]; P<.001), greater involvement of carotid territories (0.25 [IQR, 0.00–0.50] vs 0.00 [IQR, 0.00–0.25]; P=.001), and higher carotid plaque burden (2.50mm2 [IQR, 0.00–8.78] vs 1.24mm2 [IQR, 0.00–4.48]; P<.001), as well as a larger increase in total plaque area across carotid and femoral territories (7.62mm2 [IQR, 2.00–19.0] vs 5.27mm2 [IQR, 0.48–13.7]) (Table 4). Although median values appeared similar in some comparisons, the Mann–Whitney U test indicated a shift in the overall distribution of plaque progression toward higher values in the incident AO group.

After adjustment for key confounders (sex, age, systolic blood pressure, diastolic blood pressure, total cholesterol, BMI, waist circumference, smoking status, pack-years of smoking, antihypertensive treatment, statin therapy, HbA1c, and baseline atheromatous plaque), incident AO remained significantly associated with greater annual increases in total (4.40mm2; P<.001), carotid (2.33mm2; P<.001), and femoral (1.99mm2; P=.031) plaque area (Fig. 2).

Fig. 2.

Association between incident airflow obstruction and annual progression of atheromatous plaque area. Multivariable linear regression models assessed the association between Incident AO group (vs. Non-incident AO) and annual change in total atheromatous plaque area (carotid and femoral, mm2/year), with estimates and 95% robust confidence intervals (CI). Models were adjusted for sex, age, systolic and diastolic blood pressure, total cholesterol, body mass index, waist circumference, smoking status, pack-years of smoking, antihypertensive treatment, statin treatment, glycated hemoglobin (HbA1c) and baseline total atheromatous plaque area.

In sensitivity analyses without percentile trimming of annual plaque change, the results of the fully adjusted models were materially unchanged, with effect estimates of similar magnitude and direction and no meaningful differences in statistical significance compared with the primary analysis (e-Fig. 2, in supplementary material).

In sensitivity analyses using an LLN-based definition of AO (FEV1/FVC z-score<−1.645), 116 (7.6%) of 1517 participants without airflow obstruction at baseline (FEV1/FVC z-score≥−1.645) developed incident obstruction at follow-up. In fully adjusted models, effect estimates were of similar magnitude and direction compared with the primary fixed-ratio analyses (e-Table 4 and e-Fig. 3, in supplementary material).

Baseline characteristics by smoking status are presented in e-Table 5, in supplementary material. Among non-smokers, those with incident AO were older, had lower diastolic blood pressure, a distinct lipid profile, and greater statin use. In adjusted analyses, incident AO was associated with greater annual progression of total and carotid plaque in both non-smokers and smokers (e-Fig. 4, in supplementary material).