Accelerated decline in lung function is associated with cardiac structure and function and independently predicts incident heart failure

Abstract

Abstract Background The association between impaired lung function and cardiovascular disease is well known. However, it is not fully understood how a decline in lung function affects the development of cardiovascular disease including heart failure (HF). Purpose We aimed to investigate the association between declining lung function and changes in cardiac structure and function. Furthermore, we investigated the association between a decline in lung function and the risk of HF. Methods The study included 2,995 participants from a general population study. All participants had two examinations with spirometry performed 10 years apart. Participants were free of valvular heart disease and HF. A subsample (n=1,164) was also examined with echocardiography at both timepoints. The decline in the forced expiratory volume in one second (dFEV1) from the first visit to the second visit, was calculated. The change in cardiac structure and function between the two visits, was calculated as the decline in left ventricular ejection fraction (dLVEF), global longitudinal strain (dGLS), left ventricular mass index (dLVMi) and the ratio of early diastolic mitral inflow velocity to early diastolic mitral annulus velocity (dE/e’). Linear regression models were adjusted for the baseline FEV1 and echocardiographic value to account for regression to the mean. Outcome was incident HF. Results The entire sample consisted of 57% women, and the mean age was 62 ± 14 years. A total of 59% had hypertension, 8% had ischemic heart disease (IHD), 6% had diabetes, and 18% were active smokers. Mean FEV1 at baseline was 2.8 ± 0.95 liters (95 ± 19% as percentage of the predicted value). The mean decline in FEV1 between visits (median 10.8 years, IQR: 10.3-11.2) was 309 ± 362 mL. In the subsample examined with echocardiography, decreasing dLVEF was significantly associated with decreasing dFEV1 (p=0.030) in multivariable linear regression analysis, whereas dGLS, dLVMi, and dE/e’ were not. In the total study sample, 105 participants developed incident heart failure following the second visit during a median follow-up of 5.4 years (IQR: 4.5-6.3). Participants in the lowest quartile of dFEV1 (no change or increase) had an incidence rate of 2.48 per 1,000 person years (95%CI 1.33-4.60) versus participants in the highest quartile of dFEV1 (largest decline) with an incidence rate of 13.9 per 1,000 person years (95%CI 10.6-18.3), p<0.001. In multivariable Cox regression analysis, dFEV1 was an independent predictor of HF with a hazard ratio of 3.37 per 1 liter decline (95%CI 1.53-7.43, p=0.003). Conclusion We found a strong association between a decline in FEV1 and a decline in LVEF over a decade. Decreasing FEV1 was independently associated with incident HF. The incidence was highest for those with the largest decline in FEV1.A) Linear association B) Hist. of dLVEFA) Incidence rate B) Cumulative inc.