In this meta-regression analysis of 20 RCTs and 3,263 participants, we demonstrated all the changes in hemodynamic indices, including mPAP, PVR, RAP and CI were associated with ∆6MWD, independent of age, sex and baseline functional class. However, only ∆mPAP, ∆PVR and ∆CI, but not ∆RAP or ∆6MWD were related to clinical adverse events, after accounting for age, sex and functional class. In addition, only ∆PVR and ∆CI were the hemodynamic parameters to be independently predictive of total mortality. But none of the changes in hemodynamic indices was correlated with PAH hospitalization or death due to PAH. The study results may support the use of the changes in the hemodynamic parameters, including ∆mPAP, for the risk assessment in the management of PAH.
Hemodynamic indices and the prognosis
Although elevated mPAP is essential in the diagnosis of PAH and small increases in mPAP are independently associated with increased mortality in patients with borderline pulmonary hypertension, [9, 10] several studies did not demonstrate any association between mPAP and the survival in PAH patients. [4, 5] Benza et al. have shown RAP but not mPAP were associated with 1-year survival in REVEAL registry of 2,716 subjects. [11] In contrast, CI among the hemodynamic indices was suggested to be predictive of clinical outcomes in European cohorts of PAH. [3, 12] Since the difference between mPAP and PAWP is the product of cardiac output multiplied by PVR, the increase in PVR and the decrease in cardiac output along with the progression of PAH could partially cancel their effects on mPAP. Conversely, an increase in mPAP could result from an augment in cardiac output due to improving PAH, and might not be due to a raise in PVR from deteriorating PAH. Therefore, the determinants of mPAP could vary on different stages of right ventricular failure, and the prognostic value of mPAP in PAH patients is expected to be low. However, D'Alonzo et al. showed that a higher mPAP at the diagnosis of PAH conferred a greater risk of early death in a cohort of 194 patients. [13] Moreover, Sitbon et al. identified an apparently paradoxical correlation between low baseline mPAP and mortality in 178 patients with PAH in WHO functional class III or IV. [14] In patients with severe PAH and right ventricular failure, low mPAP may better correlate with low cardiac output rather than low PVR, indicating worse outcomes. [14] On the other hand, few studies have investigated the association between the changes in hemodynamic indices and outcomes in patients with PAH. Weatherald et al. presented a PAH cohort of 981 patients who had undergone repeated hemodynamic surveys in a mean time of 4.6 months. [15] The results suggested that ∆mPAP and ∆PVR were significantly associated with death or lung transplantation in the whole study population, while ∆CI was only predictive of clinical outcomes in the subgroup of severe PAH patients. [15]
In the present study, we have shown that ∆mPAP, ∆PVR, ∆RAP and ∆CI were all crudely correlated with clinical adverse events. After accounting for age, sex and WHO functional class, ∆mPAP, ∆PVR and ∆CI remained significantly related to clinical outcomes. The study results may support the inclusion of these indices in the simplified risk score for the prediction of disease outcomes. [12]
The 6-minute walk distance
The change from baseline in 6MWD has long-term served as the surrogate endpoint in the clinical trials of PAH to evaluate the therapeutic efficacy of the study drugs. It is expected the indirect measure of 6MWD may reflect the clinically meaningful endpoints, such as quality of life and survival. The SERAPHIN study may have firstly endorsed the directly clinical outcomes as the primary endpoint to demonstrate macitentan significantly reduced morbidity and mortality among patients with PAH. [16] However, the change in 6MWD was not associated with the long-term outcomes. [17] In a meta-analysis of 16 short-term RCTs, Macchia et al. have shown the change in 6MWD was not predictive of a survival benefit or adverse clinical events. [18] The updated meta-analyses have demonstrated again that the change in 6MWD didn’t correlate with any of the composite clinical events, including mortality, hospitalization for PAH, lung transplantation, or initiation of rescue therapy. [19, 20] The present study also found that ∆6MWD was not associated with clinical outcomes. The results support the use of morbidity and mortality rather than ∆6MWD as the primary endpoint in the RCTs for PAH patients.
Study limitations
The long-term prognostic values of hemodynamic changes haven’t been evaluated in large cohorts yet. Although meta-regression analysis may improve our understandings for the associations between hemodynamic indices and the long-term clinical outcomes, the variances of baseline characteristics, study designs, and background therapies across the enrolled RCTs can cause biased study findings. Some RCTs were undertaken to prove the short-term effects of a novel drug mainly on exercise capacity. Although the others might have been designed to evaluate the therapeutic effects on long-term mortality and morbidities, caution should be exercised to interpretate the correlations between hemodynamic changes and clinical outcomes. For patients with early PAH and preserved right ventricular function, the therapeutic changes in CI might be subtle, and the changes in mPAP may reflect the changes in PVR. In patients with PAH and profound right ventricular failure, improvement of CI followed by increased mPAP may indicate significant amelioration of right ventricular dysfunction and better long-term outcomes were expected. While connective tissue disease is the second common etiologies of PAH, it may cause direct damage on myocardium rather than though PAH. The inclusion of these subjects with distinct pathophysiology in the previously published RCTs may influence the findings observed in the present meta-regression analysis. Moreover, the study results are based on published RCTs, in some of which currently available PAH drugs were not commercially available. Future studies are needed to evaluate the prognostic impacts of hemodynamic indices, stratified by PAH etiologies and right ventricular function.