This prospective, single-center study was designed to determine the prognostic significance of secondary MR in HF patients with varying ePASP. We examined the records of 956 consecutive patients hospitalized with signs of HF at the Department of Cardiology, Ruijin Hospital, Shanghai Jiao Tong University of Medicine, China, from March 2017 to March 2020. Subsequently, the inclusion criteria were fulfilled by 468 patients: (1) patients with underlying heart disease that could lead to HF; (2) left ventricular dysfunction, defined as $\mathrm{˂}$50% left ventricular ejection fraction (LVEF); and (3) MR related to LV remodeling. Exclusion criteria were: (1) HF patients with intrinsic valvular deficits or prior valvular intervention, and (2) patients with noncardiac causes of PH; (3) patients with previous cardiac resynchronization therapy (CRT) or installation of implantable cardioverter-defibrillators (ICD). To exclude noncardiac causes, patients diagnosed with pulmonary arterial hypertension confirmed by right heart catheterization (RHC) either before hospitalization or during follow-up were excluded from the study cohort. Non-invasive methods such as arterial blood-gas measurement, pulmonary function tests, chest computed tomography (CT), or computed tomography pulmonary angiography were employed when there were suspicions of PH associated with lung disease or chronic thromboembolic pulmonary hypertension. The medical history was examined in order to exclude drug- and toxin-associated PH. Other biochemical, hematological, immunological, HIV, and thyroid-function tests were performed to identify other conditions related to PH. During the enrollment process, demographic data was collected, and baseline measurements including laboratory tests were conducted. Throughout the study duration, medications including renin-angiotensin-aldosterone system inhibitors and $\beta$-blockers were prescribed and titrated to reach optimal dosages for each individual with HF. During hospitalization, medication doses were gradually adjusted based on factors including blood pressure (BP), heart rates, symptoms, and laboratory test. The final dose adjustment during hospitalization occurred within 2 days before anticipated discharge. Following discharge, the patients were asked to record their daily heart rates and BP, and undergo laboratory tests. The doses of medication for HF were adjusted based on these assessments. Patients were instructed to have follow-up visits every 1–2 months during the first six months following discharge. Medication adherence was assessed through hospital visit or telephone. The follow-up interval for patients receiving optimal HF therapy was at least 3 months. CRT or ICD placement was performed during the follow-up period in accordance with current recommendations and patients’ consent [10]. Throughout the follow-up visits, a total of 6 patients underwent CRT and 12 patients had ICD installed.

The institutional review committee of Rui Jin Hospital, Shanghai Jiao Tong University of Medicine approved the study and confirmed that the research adhered to the ethical guidelines of the 1975 Declaration of Helsinki. Informed consent was obtained in writing from all participants prior to screening.

Two-dimensional and Doppler transthoracic echocardiography imaging was conducted at baseline as a routine clinical practice. All patients included in the study were hospitalized due to signs and symptoms of HF and underwent echocardiography when they were hemodynamically stable. All echocardiographic examinations were conducted using a Vivid E9 imaging system with an M5S transducer (General Electric Medical Health, Milwaukee, WI, USA), and performed by skilled cardiac sonographers. All the echocardiograms were scrutinized by cardiologists with board certification. In the parasternal long-axis view, the LV end-diastolic dimension (LVEDD), the LV end-systolic dimension (LVESD), and the left atrial dimension (LAD) were measured using two-dimensional echocardiographic images. The vena contracta width (VCW) was measured at the narrowest portion of the regurgitant flow occurring at, or immediately downstream of the orifice, using a parasternal long-axis image with magnification. The Simpson biplane approach was applied to LVEF evaluation. Using continuous-wave Doppler echocardiography we measured the peak velocity of tricuspid regurgitation. The ePASP was determined with a simplified Bernoulli equation: $\text{ePASP}=4\times {\text{peak tricuspid regurgitant velocity}}^2+\text{mean right atrial pressure}$. The mean right atrial pressure was determined based on the diameter of the inferior vena cava and variations in response to respiration. MR was quantified using the proximal-isovelocity-surface area (PISA) technique. The effective regurgitant orifice area (EROA) was obtained from the maximal PISA radius and the peak regurgitant velocity of a continuous-wave MR jet [11]. The stroke volumes were computed as the product of the time-velocity integral and the area of the annulus of each valve, using pulsed-wave Doppler methods. The regurgitant volume (RV) was computed by subtracting the stroke volumes of the mitral and aortic valves. The EROA, RV, and VCW were chosen as severity indices for MR. Severe MR was classified as an EROA $\ge$0.4 cm${}^2$, an RV $>$60 mL/beat, or a VCW $\ge$0.7 cm, whereas the non-severe MR group consisted of patients with echocardiographic indices below those levels [11]. VCW was used to classify the severity of tricuspid regurgitation (TR) in a semiquantitative manner. A VCW $\ge$0.7 cm was classified as severe TR, 0.3–0.69 cm as moderate TR, and 0.3 cm as mild TR. An ePASP of $>$50 mmHg was classified as ePASP elevation, because the value is thought to be connected to poorer results in patients with functional MR, and the prognostic effect of an ePASP of 50 mm Hg was also confirmed in our study [12].

The primary endpoint of the trial was either cardiovascular disease-related mortality or hospitalization for HF. The secondary endpoints included individual cardiovascular disease and heart failure hospitalization (HFH). Cardiovascular mortality comprised death from decompensation of HF, acute HF, myocardial infarction, fatal arrhythmia, sudden cardiac death, and thromboembolism-related deaths. HF diagnosis was based on clinical symptoms, signs, abnormal test markers, and radiologic evidence. The results were collected by reviewing medical records and/or conducting telephone or hospital-based surveys. These results were confirmed by an independent group of clinicians.

Patients were categorized based on the severity of MR: (a) non-severe MR group, and (b) severe MR group. A Kolmogorov-Smirnov test was performed for normal-distribution validation. Continuous variables in baseline characteristics were presented as the mean (SD) or median (interquartile range [IQR]) and were compared across subgroups using Student’s t-test or Wilcoxon rank-sum test, respectively. Numbers and proportions were presented for categorical variables. The Pearson chi-square test or the Fisher exact test was employed for comparison. The relationship between ePASP and the clinical outcomes was explored by using restricted cubic splines with four knots. LVEF, NT-proBNP (N-terminal pro–B-type natriuretic peptide), and LVEDD were presented as medians (IQR). We compared these three variables in the two subgroups with varying levels of ePASP with the Wilcoxon rank-sum test. The Kaplan-Meier method was utilized to describe event rates for the examination of time-to-first-event in each group with varied ePASP stages. Univariable Cox proportional hazards regression models were constructed to assess the probability of survival. The log-rank test was applied to the comparison of any two subgroups with distinct ePASP at each endpoint. Age, sex, body mass index (BMI), diabetes mellitus, hypertension, coronary artery disease, prior atrial fibrillation, hypercholesterolemia, anemia, estimated glomerular filtration rate (eGFR), and LVEF were adjusted in multivariable Cox proportional hazards regression models. The Schoenfeld residual test was used to verify the proportional hazards assumptions (Supplementary Table 1). Variance inflation factors were calculated for multicollinearity diagnosis (Supplementary Table 1). Participants who dropped from the study during the follow-up period were considered censored data. All p values were two-tailed, and p $\le$ 0.05 was judged to be statistically significant. The R package 4.2.2 (https://cran.r-project.org/) was utilized for analysis.

The study included a total of 468 consecutive patients with a primary diagnosis of HF and a secondary diagnosis of MR. Among them, 57 of 468 patients (12.1%) were identified with severe MR based on echocardiographic assessment (Fig. 1). The mean age of the cohort was 62.5 years (SD 12.0), with males accounting for 83.5% (391) of the sample. The mean LVEF was 36.8% (8.5%).

Fig. 1.

Flow-chart of recruitment, grouping and follow-up of patients with heart failure. LVEF, left ventricular ejection fraction.

Table 1 summarizes the baseline characteristics of the patients in the two subgroups. Patients with severe MR were more likely to be female, suffer from atrial fibrillation or atrial flutter and anemia, and were more likely treated with a diuretic and a mineralocorticoid-receptor antagonist (MRA). Furthermore, patients in the severe MR cohort had elevated levels of NT-proBNP and a reduced eGFR. Additionally, the majority of patients in the severe MR group had non-ischemic cardiomyopathy and were less likely to require percutaneous coronary intervention (PCI). Echocardiographic measurements revealed a higher ePASP, advanced structural enlargement, and LV dysfunction among those with severe MR. Moreover, the severe MR group demonstrated a higher incidence of TR.

A linear relationship between ePASP and risks for cardiovascular death or heart failure hospitalization is visualized in Fig. 2 (p for non-linearity = 0.64). Every 10 mmHg increase in ePASP is associated with a 32.3% increase in the risk for primary outcomes (hazard ratio [HR]: 1.32; 95% confidence interval [CI] = 1.19–1.47; p 0.001). In an ePASP of 50 mmHg, patients in the study sample are at elevated risk for cardiovascular death or heart failure hospitalization (HR: 1.49; 95% CI =1.15–1.94, p = 0.003).

Fig. 2.

Unadjusted relationship between a continuous ePASP and the hazard ratios for cardiovascular death or heart failure hospitalization in the study sample. Shades areas show the 95% CI for the HR at every single pressure level. The dashed red line represents an ePASP of 50 mmHg. ePASP, estimated pulmonary artery systolic pressure; CV, cardiovascular; HFH, heart failure hospitalization; HR, heart rates.

In the current study, a total of 100 patients (21.4% of the entire cohort) exhibited an ePASP measurement above 50 mmHg. Patients with ePASP $\ge$50 mmHg exhibited more severe LV dysfunction, as evidenced by LVEF (LVEF median [IQR]: 32.5% [IQR = 28%–40%] vs 38% [IQR = 30%–45%], p 0.001) (Fig. 3A), NT-proBNP (logarithmic form of NT-proBNP: 8.24 [IQR = 7–9] vs 6.98 [IQR = 8–9], p 0.001) (Fig. 3B), and LVEDD (LVEDD: 64 [IQR = 60–70] mm vs 61 [IQR = 57–66] mm, p 0.001) (Fig. 3C). Among patients with a ePASP $\mathrm{˂}$50 mmHg, those with severe MR exhibited poorer LVEF (35% [IQR = 29%–42%] vs 39% [IQR = 31%–45%], p 0.001), higher NT-proBNP (logarithmic form: 8.13 [IQR = 7–9] vs 6.84 [IQR = 6–8], p 0.001) and greater LVEDD (62 [IQR = 59–67] mm vs 60 [IQR = 56–65] mm, p 0.001) (Fig. 3). Similarly, severe MR was related to lower LVEF (29.5% [IQR = 27%–33%] vs 30% [IQR = 26%–34%], p = 0.01) and greater LVEDD (71 [IQR = 65–73] mm vs 66 [IQR = 59–72] mm, p 0.001) among those with higher ePASP, but there was no substantial difference in NT-proBNP between non-severe and severe MR groups (logarithmic form: 8.43 [IQR = 8–9] vs 7.96 [IQR = 7–9], p = 0.18) (Fig. 3).

Fig. 3.

The relationship between ePASP and LV function in non-severe and severe MR group. Boxplots for (A) LVEF, (B) logarithmic form of NT-proBNP, and (C) LVEDD, comparing patients with ePASP 50 mmHg and ePASP $\ge$50 mmHg in the non-severe and severe MR groups. ns, not significant; * = p 0.05; *** p 0.001. LVEF, left ventricular ejection fraction; MR, mitral regurgitation; ePASP, estimated pulmonary artery systolic pressure; NTproBNP, N-terminal pro- B-type natriuretic peptide; LVEDD, left ventricle end-diastolic dimension; LV, left ventricle.

The mean follow-up interval for the complete cohort was 578.5 (median = 381–1040) days, with 93.3% of the patients completing the follow-up. In the whole study, 146 (31.2%) of the patients met the primary endpoint, 57 (12.2%) died from cardiovascular events, and 103 (22.0%) were hospitalized for HF. The frequency of the primary endpoint was greater in the severe MR group than in the non-severe group (54.4% vs 28.0%; p 0.0001; Fig. 4A). Patients with severe MR exhibited higher risk for a composite outcome of cardiovascular mortality and HFH (HR = 1.88; 95% CI = 1.42–2.31; p 0.001, and adjusted hazard ratio (AHR) = 1.40; 95% CI = 1.03–1.91; p = 0.03) (Table 2). The incidence of cardiovascular death (Fig. 4B) and HFH (Fig. 4C) were also greater in the severe MR group. Patients with severe MR had higher risk for cardiovascular mortality (AHR = 1.82; 95% CI = 1.15–2.88; p = 0.011; Supplementary Table 2). HFH was comparable in the two groups (AHR = 1.17; 95% CI = 0.78–1.77; p = 0.44; Supplementary Table 3).

Fig. 4.

Accumulated risks for both the primary and secondary endpoints in all patients classified by degrees of MR. Survival curves for the primary endpoint (A) (log rank p 0.0001), cardiovascular death (B) (log rank p 0.0001) and HFH (C) (log rank p = 0.0081) comparing patients with in non-severe and severe MR groups. Log rank test was used to compared the difference between two groups. CV, cardiovascular; HFH, heart failure hospitalization; MR, mitral regurgitation.

The clinical outcome of patients was evaluated with different ePASP levels and stratified by baseline MR severity. Among patients with ePASP 50 mmHg there was no significant difference for the primary endpoint in the severe MR group (42.9% vs 25.5%; p = 0.051; Fig. 5A). There was a non-significant “trend” toward increased risk for the primary endpoints in the severe MR group (HR = 1.71; 95% CI = 0.99–2.97; p = 0.055; Table 3), but the differences were not statistically significant at this sample size. This “trend”, however, was no longer evident after full adjustments (AHR = 0.99; 95% CI = 0.54–1.84; p = 0.98; Table 3). Fig. 6A depicts the frequency of primary endpoints in patients with ePASP $\ge$50 mmHg (72.7% vs 38.5%; p 0.0001). Severe MR was associated with greater risk for primary endpoints (HR = 3.62; 95% CI = 1.96–6.70; p 0.001; Table 4), and the difference remained significant after adjustments (HR = 2.76; 95% CI = 1.28–5.98; p = 0.0098; Table 4).

Fig. 5.

Accumulated risk for both the primary and secondary endpoints in patients with ePASP 50 mmHg, classified by degrees of MR. Survival curves comparing patients in non-severe and severe MR groups for (A) the primary endpoint (log rank p = 0.051); (B) cardiovascular death (log rank p = 0.047); and (C) HFH (log rank p = 0.089). CV, cardiovascular; HFH, heart failure hospitalization; MR, mitral regurgitation; ePASP, estimated pulmonary artery systolic pressure.

Fig. 6.

Accumulated risks for both the primary and secondary endpoints in patients with ePASP $\mathrm{\ge }$50 mmHg, classified by degrees of MR. Survival curves comparing patients in non-severe and severe MR groups for (A) the primary endpoint (log rank p 0.0001); (B) cardiovascular death (log rank p 0.0001; and (C) HFH (log rank p = 0.063). CV, cardiovascular; HFH, heart failure hospitalization; MR, mitral regurgitation; ePASP, estimated pulmonary artery systolic pressure.

Patients within the severe MR group with ePASP 50 mmHg exhibited a higher incidence cardiovascular death (Fig. 5B). However, there was no significant difference in cardiovascular mortality risk between the two groups (AHR = 1.34; 95% CI = 0.52–3.48; p = 0.54; Supplementary Table 4). With respect to patients with ePASP $\ge$50 mmHg, the frequency of cardiovascular death was considerably greater in the severe MR group than in the non-severe group (Fig. 5B). Patients with severe MR also had an elevated risk of cardiovascular death (AHR = 5.50; 95% CI = 1.88–16.05; p = 0.0018; Supplementary Table 5) (Fig. 6B). Hospitalization rates and hospital risk for HF were comparable between the two groups regardless of ePASP (Figs. 5C,6C; Supplementary Tables 6,7).