The objective of this study was to investigate the prevalence, characteristics and survival of pulmonary hypertension due to chronic heart failure, with HFrEF and HFpEF analyzed separately, in a prospective multicenter setting. We demonstrated that: 1) the prevalence of PH was 67.2% in HFrEF and 40.2% in HFpEF, respectively; 2) patients with PH-LHD presented significantly worse echocardiographic and invasive hemodynamic parameters regardless of HF etiology; 3) predictors of PH differed between PH etiologies, but LVEDD could be a consistent predictor of PH both in HFrEF and in HFpEF; 4) the survival of PH patients was significantly worse than patients without PH; and 5) DPG was a significant prognostic variable both in HFrEF and HFpEF.
The true prevalence of PH due to HF is unknown, which can be attributed to the heterogeneity of epidemiological studies based on different study designs, study cohorts, diagnostic modalities, or various definitions and thresholds of PH. It is estimated that the prevalence of PH-LHD in HF ranges from 25–83%[7–15].The prevalence of PH-LHD reported in this study is accordance with what have reported before, indicating PH is a common complication in HF regardless of etiology. However, it should be noted that, compared to Group I PH, RHC was not common or of routine practice in patients with heart failure. Usually, catheterization is performed for suspected PH or for evaluation prior to surgical procedures such as transplantations. It is also true in our institutes, which implies the actual prevalence of PH can be overestimated in literatures and in our study as not all HF patients would undergo RHC in daily practice. It can also explain why we only enrolled 500 patients in our study period despite the large HF population in China[27]. The relatively small sample size could be attributed to our inclusion criteria, as we only enrolled patients undergoing RHC.
It should be noted that ischemic heart disease accounted for the most prevalent etiology in our study (77.2%), and RHC was combined with a left heart catheterization for coronary angiography in 85.6% of all patients, whose PAWPs were substituted by LVEDPs. It is recognized that when measured properly, the PAWPs should closely approximate LVEDPs[26, 28]. However, the accurate measurement of PAWPs has been a challenge in many clinical circumstances, such as the respiratory status of measurement, the existence of a larger left atrial diameter and atrial fibrillation[28]. Therefore, when LVEDPs could be attained in our study, we used LVEDPs to classify PH-LHD.
In our cohort, males accounted for 74.6% of the cohort, with a similar distribution between HFrEF and HFpEF, which was higher than the data reported by China-HF Registry[27], and opposite to the knowledge that HFpEF is thought to be more prevalent in women[8]. The paucity of women with this condition could be explained by the inherent sex inequality in the patient population, as the majority of the patients in the study was patients with coronary artery disease. Besides, It is reported that in heart failure treatment, barriers exist for seeking optimal quality care between men and women, with women less frequently receiving advanced treatments[29, 30]. It is also noteworthy that we cannot ignore the influence of sex inequality on the results reported by this study.
Besides, our data showed that patients with PH had worse hemodynamics and a significant worse survival than patients without PH. However, although the adverse effects of PH on the prognosis have been more and more recognized, the management of PH-LHD is still restricted to the treatments of HF, without sufficient evidence on the benefit of PH-targeted drugs in this patient population[26]. Nevertheless, as we reported, a large proportion of patients in our study were not on optimal therapy of HF, evidenced by the low usage of HF medications. The data was similar to China-HF Registry[27], reporting the usage of angiotensin-converting enzyme inhibitors/angiotensin receptor blockers (ACEIs/ARBs),β-Blockers and mineralocorticoid receptor antagonists (MRAs) were more common in HFrEF than HFpEF, and less implemented in China compared to the data of HF registries in developed countries, such as ATTEND, ADHERE and EHFS Ⅱ[31–33]. Our data has provided insights into the gaps in clinical practice of the HF management in China, indicating efforts are needed to improve the compliance to recommended guidelines.
Various risk factors, from clinical presentation to parameters obtained from echocardiography or other tests, have been reported to be related to the presence of PH-LHD[4]. Despite several risk stratification strategies has been proposed to help differentiate PH-LHD from pre-capillary PH[22–25], predictors of PH in HF patients still lack demonstrations. In a HFrEF cohort, Miller et al. [21]found that the presence of PH was associated with older age, diuretic use, atrial fibrillation, and lower PAC. Besides, Gerges et al.[9] demonstrated that chronic obstructive pulmonary disease and the tricuspid annular plane systolic excursion to systolic pulmonary artery pressure ratio predicted combined post- and pre-capillary PH (Cpc-PH) in HFrEF, while in HFpEF, younger age, valvular heart disease, and the tricuspid annular plane systolic excursion to systolic pulmonary artery pressure ratio predicted Cpc-PH. In our cohort, BMI and LVEDD were found to be predictors of PH in HFrEF, and the female gender, LVEDD and RVAPD were associated with the presence of PH in HFpEF. As we discussed above, though as the gold standard to diagnose PH, RHC remains less utilized in HF. Consequently, demographics, clinical presentation and results of noninvasive tests, especially echocardiography, are still more helpful references to identify PH in HF population. As none of risk factors can consistently predict PH in different study cohorts, a comprehensive evaluation is needed for every patient with their relevant data fully considered. Meanwhile, further studies are still required to investigate how different HF etiologies, including hypertensive heart disease, ischemic heart disease or cardiomyopathy would affect the hemodynamic phenotypes of PH-LHD due to HF.
Till now, none of the hemodynamic parameters are free from limitations. The most debated point of DPG is its conflicted prognostic value[9]. Therefore, PVR, considered as a better reflection of pre-capillary component, was then reintroduced to define Cpc-PH incorporating with DPG[8]. However, reported by Gerges M et al.[9], PVR alone was a predictor of outcome only in HFrEF but not in HFpEF, while DPG retained significant in both etiologies. In our cohort, we found increasing DPG was a significant predictor for mortality, and patients whose DPG ≥ 7 mmHg had a worse survival both in HFrEF and in HFpEF, which can provide extra prognostic information about the hemodynamics of PH-LHD in a prospective multicenter study setting. The optimal combination of variables is still under debate, and it could be better to use those variables in combination rather than focus on an isolated value[26].
Our study has several limitations. First, the proportion of patients who met primary endpoint was relatively small, which might had a potential effect on the differentiation power regarding the comparisons of survival curves, as well as the Cox regression analyses. In addition, data about right cardiac function, or right ventricular to pulmonary vascular coupling were not available in the present study. Besides, as only baseline characteristics were included in the analysis, with the absence of follow-up data, further studies are needed to explore the the effect of changing hemodynamics and medical interventions.