Baseline characteristics

Mean age of the study population was 50±17 years; most patients were female (73%), and the majority of patients were diagnosed with idiopathic pulmonary arterial hypertension (59%; online supplemental table 1). Most patients were in New York Heart Association functional class III (50%), the mean 6 min walking distance was 411±145 m and N-terminal pro-B-type natriuretic peptide (NT-proBNP) levels were 1360 ng/L (358–2625). Patients presented with an mPAP of 55±14 mm Hg, mixed venous oxygen saturation of 63%±9%, mean right atrial pressure (mRAP) of 8±5 mm Hg and PVR of 843±384 dynes/sec/cm−5. CMR imaging showed RV dilatation reflected by an indexed RVEDV of 83±21 mL/m2 and indexed RVESV of 55±21 mL/m2 resulting in a mean RVEF of 35%±12% at initial presentation. An overview of disease severity is presented in online supplemental table 2.

RVEF≥45%@FU patients versus RVEF<45%@FU patients

The first follow-up was performed after a median time of 1.0 year (0.9–1.1). At first follow-up, 75 patients (59%) reached an RVEF of 45% or maintained above this threshold (RVEF≥45%@FU patients), whereas in 52 patients (41%) RVEF remained below 45% (RVEF<45%@FU patients; figure 1). Three out of 29 patients failed to maintain their RVEF above the 45% threshold. RVEF≥45%@FU patients had a less severe RV phenotype at baseline, characterised by greater RVEF (p<0.001), moderate RV dilatation (77±18 mL/m2 vs 91±21 mL/m2; p<0.001) and lower PVR (785±397 dynes/sec/cm−5 vs 947±341 dynes/sec/cm−5; p=0.034). There were no differences in gender and age between RVEF<45%@FU and RVEF≥45%@FU patients (online supplemental table 3). Baseline multivariable predictors of an RVEF≥45%@FU are RVEDVi (OR (95% CI): 0.96 (0.94 to 0.98), p<0.001) and NT-proBNP (OR (95% CI): 0.999 (0.9992 to 0.0.9998), p<0.001) (online supplemental table 4).

Figure 1

Longitudinal visualisation of RVEF stratified for patients who reach or maintain the 45% threshold. P-values represent the change over time within each group. The red dotted line indicates the 45% cut-off. Dots and whiskers represent the mean±SEM. CMR, cardiac magnetic resonance; FU, follow-up; RVEF, right ventricular ejection fraction; RVEF≥45%@FU, patients with an RVEF ≥45% at first FU; RVEF<45%@FU, patients with an RVEF <45% at first FU.

Haemodynamic determinants

At first follow-up, both groups showed haemodynamic and functional improvement. An overview is presented in online supplemental table 3. PVR, mRAP and mPAP significantly decreased in both groups. However, the magnitude of mPAP decrease was larger in RVEF≥45%@FU patients (figure 2A; pinteraction <0.001). Also, RVEF≥45%@FU patients showed a greater decrease in PVR (figure 2B; pinteraction=0.04). Both groups had a similar improvement in CI (pinteraction=0.37). Naturally, RVEF improvement was greater in RVEF≥45%@FU patients (pinteraction<0.001), as well as SV; greater reductions in RV remodelling were observed in RVEF≥45%@FU patients for RVEDVi (pinteraction<0.01), RVESVi (pinteraction<0.001) and RV mass (pinteraction<0.001) than in RVEF<45%@FU patients. This was reflected in a greater reduction of NT-proBNP levels for RVEF≥45%@FU patients (pinteraction=0.03). No differences in LV function improvement could be observed between RVEF≥45%@FU patients and RVEF<45%@FU patients.

Figure 2

Changes in mPAP and PVR as stratified by an RVEF of 45% @ FU. (A) Longitudinal visualisation of mPAP in RVEF≥45%@FU patients and RVEF<45%@FU patients. (B) Longitudinal visualisation of PVR in RVEF≥45%@FU patients and RVEF<45%@FU patients. P-values represent the change over time within each group by two-way repeated measures analysis of variance. Dots and whiskers represent the mean±SEM. FU, first follow-up; mPAP, mean pulmonary arterial pressure; PVR, pulmonary vascular resistance; RHC, right heart catheterisation; RVEF, right ventricular ejection fraction; RVEF≥45%@FU patients, patients with a right ventricular ejection fraction ≥45% at first FU; RVEF<45%@FU patients, patients with an RVEF<45% at first FU.

RV phenotype

RVEF≥45%@FU patients were characterised by less severe RV dilatation at initial presentation compared with RVEF<45%@FU patients. Remarkably, improvement of RV volume and function can be achieved, even in patients with severely dilated right ventricles. Figure 3 displays the improvement of RVEF from baseline to follow-up, stratified by RVEDVi and RVESVi quintiles. The proportion of patients reaching the 45% RVEF threshold declines when the RV is more dilated at baseline. Haemodynamic profiles between RVEDVi quintiles were comparable (online supplemental table 5, RVESVi quintiles; online supplemental table 6).

Figure 3

Absolute change in RVEF for baseline RVEDVi and RVESVi quintiles. (A) Absolute change in RVEF for baseline RVEDVi subdivided into quintiles. (B) Absolute change in RVEF for baseline RVESVi subdivided into quintiles. P-values represent significant change in RVEF for each RVEDVi and RVESVi quintile determined through pairwise t-tests. Values above the x-axis represent portion of patients reaching the 45% RVEF threshold. The red dotted line indicates the 45% cut-off. Dots and whiskers represent the mean±SEM. CMR, cardiac magnetic resonance; FU, first follow-up; RVEDVi, right ventricular end-diastolic volume index; RVEF, right ventricular ejection fraction; RVESVi, right ventricular end-systolic volume index.

Pulmonary vascular resistance and pulmonary arterial pressure

We divided our study population into quintiles based on relative PVR change from baseline to follow-up to assess the corresponding RVEF improvement. In all quintiles, individual improvement of RVEF is seen. However, the proportion of patients reaching the 45% RVEF threshold was lower when PVR reduction is smaller (online supplemental figure 2A). Additionally, we tested the hypothesis of 40% relative PVR reduction in restoring RV function. Exceeding 40% PVR reduction improves RV function in near all patients, but does not guarantee an RVEF≥45%@FU. Of interest, 56 patients had a reduction in PVR of more than 40% at follow-up. In almost all patients (53/56), RV function improved significantly (table 1). As stated previously, RVEF≥45%@FU patients also have greater reduction in mPAP. A similar approach to assess the corresponding RVEF change within each relative quintile of mPAP reduction was used (online supplemental figure 2B). A greater reduction in pulmonary artery pressure resulted in an increased proportion of patients reaching the 45% RVEF threshold.

Table 1

Change in RVEF for PVR reduction of 40%

Long-term follow-up

During follow-up, 45 (36%) patients died and 10 (7%) patients underwent lung transplantation. The second follow-up was performed after 2.1 years (1.9–3.0) and the third follow-up after 3.3 years (2.8–5.2). Between the first and second follow-up, 2 patients received a lung transplantation and 26 patients died. Between the second and third follow-up, 4 patients received a lung transplantation and 11 patients died. Importantly, reaching or maintaining an RVEF of at least 45% at first follow-up was associated with better 1-year, 3-year and 5-year survival rates. The survival rates after first follow-up were 96%, 87% and 74% for the RVEF≥45%@FU patients and 92%, 67% and 63% for RVEF<45%@FU patients (p=0.018, HR: 1.91 (95% CI: 1.11 to 3.27), figure 4). At first follow-up, 15 (20%) RVEF≥45%@FU patients received treatment intensification compared with 16 (31%) RVEF<45%@FU patients. A significantly different RVEF trajectory was seen between RVEF≥45%@FU patients and RVEF<45%@FU patients. Furthermore, reaching or maintaining a 45% RVEF at first follow-up was associated with a stable RVEF over the next consecutive visits. Patients failing to reach or maintain this 45% RVEF threshold at first follow-up, mostly stayed below this threshold (figure 5).

Figure 4

Survival stratified for patients with RVEF≥45%@FU compared with RVEF<45%@FU patients. FU, first follow-up; PAH, pulmonary arterial hypertension; RVEF, right ventricular ejection fraction. Log-rank test shows a significant different survival between both groups starting from first follow-up (p=0.018).

Figure 5

Longitudinal visualisation of RVEF in RVEF≥45%@FU patients and RVEF<45%@FU patients. (A) RVEF trajectory during follow-up for RVEF≥45%@FU patients and RVEF<45%@FU patients at two consecutive follow-up visits. (B) RVEF trajectory during follow-up for RVEF≥45%@FU patients and RVEF<45%@FU patients at three consecutive follow-up visits. The red dotted line indicates the 45% cut-off. Dots and whiskers represent the mean±SEM. CMR, cardiac magnetic resonance; FU 1, first follow-up visit; FU 2, second follow-up visit; FU 3, third follow-up visit; RVEF, right ventricular ejection fraction; RVEF≥45%@FU, patients with a right ventricular ejection fraction ≥45% at first FU; RVEF<45%@FU, patients with an RVEF<45% at first FU. Repeated measures analysis of variance shows a significant different trajectory between both groups (p<0.001).

Clinical events

An event of death or lung transplantation occurred in 23/75 (31%) RVEF≥45%@FU patients compared with 32/50 (64%) RVEF<45%@FU patients. Although survival rates were in favour of the RVEF≥45%@FU patients, a substantial amount still deceased or was listed for lung transplantation. In RVEF≥45%@FU patients, only 9/23 (39%) experienced a clinical event due to progressive RV failure. In contrast, a clinical event in 23/32 (72%) RVEF<45%@FU patients was caused by progressive RV failure. Additionally, the mean age of diagnosis in RVEF≥45%@FU patients experiencing a clinical event was 60±19 years vs 50±15 years, resulting in less lung transplantations. Data for each individual experiencing a clinical event are summarised in online supplemental table 7 (RVEF≥45%@FU patients) and online supplemental table 8 (RVEF<45%@FU patients).