Prognostic value of echocardiographic changes in patients with pulmonary arterial hypertension receiving parenteral prostacyclin therapy

Abstract

Background

It is unknown whether the echocardiographic changes observed after treatment of pulmonary arterial hypertension (PAH) patients have prognostic value.

Methods

We retrospectively identified subjects with PAH, confirmed by right heart catheterization, who had Doppler echocardiograms before (baseline) and after a year of treatment (follow-up) with parenteral prostacyclin analogues. Echocardiographic parameters were measured off-line by two investigators.

Results

We included 48 patients, age 45 ± 14 years and 83 % women. When compared to the baseline, the follow-up echocardiogram showed a reduction in the right atrial area (12 ± 25 %, p < 0.001), right ventricular (RV) basal and mid-cavity dimensions (8.5 ± 14 %, p<0.001 and 6.8 ± 17 %, p=0.005, respectively) and peak tricuspid regurgitant velocity (10 ± 14 %, p<0.001). Tricuspid Annular Plane Systolic Excursion (TAPSE) (36 ± 43 %, p<0.001) and the RV outflow flow time-velocity integral (48 ± 66 %, p < 0.001) increased. During a median (interquartile range) follow up of 52.5 (20.5 – 80) months, 18 patients (37.5 %) died, mostly (n=15, 83 %) from PAH progression. The change in RV end-diastolic area (hazard ratio (HR) per 10 % decrease: 0.73 (95% CI: 0.57–0.93)), tricuspid valve regurgitation velocity (HR per 10 cm/s decrease: 0.58 (95% CI: 0.37–0.89)), RV outflow tract velocity-time integral (HR per 10% increase: 0.90 (95% CI: 0.83–0.98)) and subjective RV function (HR per 1 unit of improvement [e.g. moderate to mild]: 0.55 (95% CI: 0.31–0.96)) were associated with overall mortality.

Conclusions

Echocardiographic parameters that estimate right ventricular systolic pressure and assess RV morphology and function improve after a year of prostacyclin analogue treatment and the degree of change has prognostic implications.

Introduction

Pulmonary arterial hypertension (PAH) is a severe and progressive pulmonary vascular condition due to narrowing of the blood vessels in the lung that can lead to right heart failure and premature death [1]. Nine pharmacological agents are FDA approved for the treatment of PAH, including two parenteral prostacyclin analogues: epoprostenol (intravenous) and treprostinil (intravenous and subcutaneous). These two parenteral therapies improved symptoms, exercise capacity and hemodynamic status in patients with PAH [2–5]. Furthermore, treatment with epoprostenol was associated with a survival benefit in patients with idiopathic PAH [2]. Notwithstanding major therapeutic advances in the last decade, the morbidity and mortality of PAH continue to be unacceptably high with a 1-, 3 and 5-year survival rates from the time of diagnosis of 85, 68 and 57 %, respectively [6].

Echocardiography is a well-established and widely available technique that is routinely used during the initial assessment of patients suspected of having PAH. In addition, it continues to be an important method to follow patients with PAH and evaluate their treatment response [7]. This noninvasive technology permits the serial evaluation of right ventricular (RV) size and function as well as the estimated RV systolic pressure. Over the years, limited attention has been paid to the effects of PAH-specific treatment on echocardiographic parameters and their prognostic implications [8].

Although a few echocardiographic parameters are known to improve after the initiation of PAH-specific therapies, it is not clear whether these changes have prognostic value. We hypothesized that improvements in certain echocardiographic parameters are associated with longer survival. We sought to investigate the effect of a year of treatment with parenteral prostacyclin analogues on echocardiographic parameters. Our main study objective is to determine whether changes in echocardiographic parameters resulting from advanced PHA-specific therapies have prognostic significance.

Methods

a) Study design and inclusion/exclusion criteria

This retrospective study was approved by the Cleveland Clinic Institutional Review Board (protocol number 10-1127). We identified eligible subjects by using the Cleveland Clinic Pulmonary Hypertension Registry. All patients had the diagnosis of precapillary pulmonary hypertension confirmed by right heart catheterization (mean pulmonary artery pressure ≥ 25 mm Hg with a pulmonary artery occlusion pressure ≤ 15 mm Hg). During RHC, patients were supine in a steady state, relaxed, and breathing room air or oxygen to maintain pulse oximetry > 90%. Patient received no sedation. We cannulated preferably the right internal jugular vein with using minimal local anesthesia (lidocaine 2 %). We zeroed pressure transducers at fourth intercostal space of the midaxillary line. Cardiac output was determined by thermodilution and Fick methodology. The resting oxygen consumption (mL/min) for the Fick equation was estimated by the formula of Dehmer et al.[9] whereas oxygen consumption=125 × body surface area. Body surface area was calculated according to the formula of Dubois [10, 11]. Mixed venous oxygenation was measured in the blood obtained from the pulmonary artery during the right heart catheterization.

We identified 112 consecutive patients who received parenteral prostacyclin analogues for at least a year, from 1/1/2004 until 1/8/2011. We selected the initiation date (1/1/2004) based on the availability of syngo Dynamics (Siemens Medical Solutions USA, Inc., Malvern, PA, USA), a system that allows for off-line echocardiographic measurements.

Each patient underwent a thorough clinical evaluation to identify the cause of PH. We excluded patients who had pulmonary hypertension other than group I as defined by the 4^th World Symposium on Pulmonary Hypertension [12] (chronic thromboembolic pulmonary hypertension (n=10) and pulmonary hypertension due to sarcoidosis (n=4). We also excluded patients with complex congenital heart diseases or those in whom surgically correction was performed in the two years prior to initiation of the study (n=5), moderate or severe mitral and/or aortic valve stenosis and/or insufficiency (n=16), echocardiograms either not done in the predetermined time window (10) or not available for off-line review (studies done at an outside hospital or not retrievable (n=19)). Patients with PH due to congenital intracardiac shunts were not excluded. No patient had evidence of left ventricular myocardial disease.

b) Measurements and calculations

Transthoracic Doppler echocardiography was performed during the initial evaluation (before the initiation of parenteral prostacyclin analogues) and at the end of a 1-year treatment with parenteral prostacyclin analogues. We arbitrarily chose a year of treatment to assess patients on an adequate and stable dose of prostacyclin analogue and allow enough time for this medication to exert a noticeable effect. Studies were performed following American Society of Echocardiography guidelines [13, 14]. All recordings were reviewed with our offline quantification system (syngo Dynamics). Measurements were obtained by experienced physicians blinded to the patients’ clinical history or survival status. All new determinations were compared with the ones provided in the report at the time of echocardiography. In case of discrepancies another physician reviewed the echocardiographic images and a consensus was obtained.

In all but three echocardiograms the patients were in sinus rhythm during image acquisition. In three studies, patients had atrial fibrillation with well controlled heart rates (two patients during the initial and one during the 1-year echocardiogram), therefore measurements were made on three beats and the results averaged. Left atrial area was measured by planimetry at the end-ventricular systole on the apical 4-chamber view. We measured RV basal, mid cavity and longitudinal dimensions of the RV on the apical 4-chamber view. In addition, RV area was obtained at end-diastole by tracing the endocardial border in the apical 4-chamber view and including trabeculations, tricuspid leaflets and chords as part of the chamber [14]. The intraclass correlation coefficient (95% confidence interval) for RV end-diastolic area determination (n=15) for the same and different raters was 0.95 (0.85–0.98) and 0.93 (0.81–0.97), respectively. Abnormal left ventricular diastolic function was divided in 3 grades, following recommendations from the American Society and European Association of Echocardiography (grade I: impaired LV relaxation, grade II: pseudo normal LV filling and grade III: restrictive LV filling) [15]. RV systolic function was visually estimated as normal, mild, moderate or severe by experienced operators. The severity of the tricuspid regurgitation was graded as mild, moderate or severe following recommendations of the American Society of Echocardiography [16]. Pericardial effusion was evaluated by 2D echocardiography at end-diastole and graded as trace (separation of pericardial layers only in systole), small (separation < 1 cm), moderate (separation ≥1 cm but <2 cm), and large (separation ≥2 cm) [17].

The RV outflow acceleration time was measured between the onset and the maximal velocity of the pulsed-wave Doppler flow profile. In addition we determined the time-velocity integral of the RV outflow tract and noted the presence of absence of midsystolic notching. Maximal tricuspid regurgitant jet velocity was obtained during quiet respiration after analyzing the continuous-wave Doppler from different echocardiographic views. In cases of noticeable respiratory spectral oscillations, measurements were obtained during brief held-expiration or three consecutive signals were averaged. We estimated the pulmonary vascular resistance (PVR) by echocardiography using the model proposed by Opotowsky et al. (PVR = (pulmonary artery systolic pressure/RV outflow tract velocity-time integral) + 3 if RV outflow tract notch is present ) [18]. For all the echocardiographic measurements, absolute change was calculated by subtracting the determination performed at 1-year of treatment from the baseline value. Percentage change was obtained by dividing the absolute change over the baseline value and multiplying by 100.

c) Follow-up

Patients were followed in our clinic at least every 3 months after the initiation of parenteral prostacyclin analogues. Echocardiography was performed before the initiation of PAH-specific therapies and every 3 to 6 months. After a year of treatment patients were continued on parenteral prostacyclin analogues unless they underwent lung transplantation or died. In no patient, prostacyclin analogue was discontinued based on echocardiographic deterioration or lack of improvement. Patients were transplanted in the event of refractory PAH, using criteria suggested by the International Society for Heart and Lung Transplantation [19]. Death of the study participants was ascertained by reviewing our records and querying the U.S. Social Security Death Index.

Statistics

Means with standard deviations (SD) and number of patients with percentages (% ) are provided for continuous and categorical variables, respectively. Comparison of echocardiographic variables at baseline and after 1-year of treatment was performed using McNemar or paired student t-test, as appropriate. Inter-rater and intra-rater agreement for single measures were calculated using the intraclass correlation coefficients and their respective 95% confidence intervals. Each subject was rated by the same two raters. We tested for absolute agreement as systematic differences are considered relevant. Survival at each time point was assessed by Kaplan-Meier methodology. The start point was the date of the echocardiogram obtained after a year of parenteral treatment. The end of follow-up was marked by the patient’s death. Patients were censored either at the time of lung transplantation or the end of the study in May, 2013. Cox proportional hazards modeling adjusted by age and gender was used to examine the relationship between survival and selected echocardiographic variables. The results are expressed as hazard ratios (HR) and the corresponding 95% confidence interval (95% CI). We generated models testing three outcome variables (overall mortality, PAH-associated death or the combination of death and lung transplantation). Predictors with a HR less than 1 are associated with lower risk for the outcome tested. As an example a HR of 0.73 means that the outcome of interest (i.e. mortality) decreases by a factor of 0.73 (27 % less) for each specified unit change of the predictor (e.g. 10 % decrease in RV end-diastolic area). For a 20 % decrease in RV area the HR for the outcome decreases by a factor of (0.73)²=0.53 (47% lower risk of having the outcome).

Receiver operating characteristic curve (ROC) was used to determine the sensitivity and specificity of different cut-offs of the change in tricuspid regurgitation velocity to discriminate patients that die during follow-up. All p values reported are two-tailed. A p value of < 0.05 was considered significant. The statistical analyses were performed using the statistical package SPSS, version 17 (SPSS Inc; Chicago, IL).

Results

1- Overall characteristics of the patients

We included at total of 48 patients (table 1) with PAH of whom 32 (67%) had either idiopathic (n=25, 52 %) or heritable (n=7, 15 %) PAH. A few patients had Eisenmenger syndrome due to ventricular septal defect (n=2) and atrial septal defect with anomalous pulmonary venous return (n=1). Six-minute walk test was obtained the same day of the echocardiogram. Right heart catheterization was done within a month of the first echocardiogram in 39 (81 %) patients.

Table 1.

Patient characteristics immediately before the initiation of parental prostacyclin analogues.

	Mean ± SD, n (%)
Number of patients	48
Age (years)	44 ± 14
Female gender	40 (83 %)
Caucasian race	40 (83 %)
Cause of PAH
-Idiopathic/heritable	32 (67%)
-Connective tissue disease	10 (21 %)
-Porto-pulmonary	3 (6 %)
-Congenital heart diseases	3 (6 %)
NYHA*
-III	24 (57 %)
-IV	18 (43 %)
6MWT distance walked (m)	317 ± 107
6MWT distance walked (% of predicted) [40]	54 ± 17
Hemodynamics
-RA pressure (mmHg)	12 ± 7
-Mean PAP (mmHg)	54 ± 12
-PAOP (mmHg)	11 ± 5
-CO thermodilution (L/min)	4 ± 1
-CO by FICK method (L/min)¶	4 ± 1
-PVR (Wood Units)	13 ± 6
-Mixed venous oxygenation (%)	60 ± 9

Abbreviations: 6MWT: six-minute walk test, NYHA: New York Heart Association, CO: cardiac output, PAH: pulmonary arterial hypertension, PAOP: pulmonary artery occlusion pressure, PAP: pulmonary artery pressure, RA: right atrium.

^*New York Heart Association (NYHA) functional class at the time of the initial echocardiogram was available in 42 patients.

^¶Oxygen consumption was estimated by the formula of Dehmer et al. [9].

2- Prostacyclin analogue treatment

All patients were treated with parenteral prostacyclin analogues for at least one year. The prostacyclin analogues used during this period were IV epoprostenol: 42 (88 %), IV treprostinil: 3 (6 %), SQ treprostinil: 2 (4 %). One (2%) patient was converted from IV epoprostenol to IV treprostinil during the first year of treatment. Twenty-five (52%) patients were receiving other PAH-specific therapies before the initiation of prostacyclin analogues (endothelin receptor antagonists (ERA): 17 (68 %), phosphodiesterase-5 inhibitors (PDE-5 inh): 3 (12 %), combination of ERA and PDE-5 inh: 5 (20 %)). One patient was initiated on a PDE-5 inh during the first year of prostacyclin analogues.

3- Serial echocardiographic determinations

We analyzed the initial echocardiogram and an echocardiogram performed after a year of treatment with parenteral prostacyclin analogues (Figure 1). The median (interquartile range: IQR) time between these two echocardiograms was 12.9 (11–14.8) months. Significant echocardiographic differences between studies reflected an increase in left sided cardiac chambers, a reduction on the right sided heart cavities, an improvement in left and right ventricular functions and a reduction in the leftward shifting of the interventricular septum (IVS) (table 2). In the echocardiogram, obtained after a year of prostacyclin analogue treatment, the peak tricuspid regurgitant velocity, estimated right ventricular systolic pressure, ratio of tricuspid regurgitant velocity/RV outflow tract time-velocity integral, estimated PVR, percentage of studies showing right ventricular outflow tract notching and grade of left ventricular diastolic dysfunction decreased, meanwhile, the RV outflow tract flow acceleration time increased (table 3). Non-significant echocardiographic parameters are shown in e-table 1.

Figure 1. Echocardiograms at baseline and after 1 year of treatment with prostacyclin analogue.

RV dimensions (panel A), tricuspid regurgitant jet (panel B) and RV outflow tract flow (panel C) of the baseline echocardiogram. RV basal, mid-cavity and longitudinal dimensions are 5.7, 4.8 and 9.2 cm, respectively. The tricuspid regurgitation velocity is 4.1 m/s. The RV outflow tract flow has a mid-systolic notch (arrow) and the velocity-time integral is 7.4 cm.

RV dimensions (panel D), tricuspid regurgitant jet (panel E) and RV outflow tract flow (panel F) of the echocardiogram after 1 year of treatment. RV basal, mid-cavity and longitudinal dimensions are 5, 4.5 and 9.6 cm, respectively. The tricuspid regurgitation velocity is 3.3 m/s. The RV outflow tract flow has no notch and the velocity-time integral is 13.1 cm.

Table 2.

M-mode and 2-D echocardiographic determinations before and after a year of parenteral prostacyclin analogue treatment in patients with PAH.

	n	Initial echocardiogram Mean ± SD, n (%)	Echocardiogra m after 1 year of treatment Mean ± SD, n (%)	Difference between studies Mean ± SD	Difference between studies		Percentage change Mean ± SD	p (paired student t- or McNemar test)
					Lower 95% CI	Upper 95% CI
Heart rate (bpm)	45	87.7 ± 15	81.9 ± 10	5.8 ± 18	0.5	11.2	3.3 ± 26	0.03
Body surface area (kg/m2)	48	1.9 ± 0.3	1.9 ± 0.3	0.02 ±0.1	−0.02	0.06	0.4 ± 7.4	0.38
Left atrial area (cm2)	45	14.4 ± 5	16.2 ± 4	1.7 ± 4	0.4	3.1	20 ± 39	0.01
Left ventricular end-diastolic diameter (cm)	44	3.3 ± 0.6	4 ± 0.7	0.7 ± 0.7	0.5	0.9	24 ± 27	< 0.001
Left ventricular end-systolic diameter (cm)	44	2.1 ± 0.5	2.5 ± 0.6	0.4 ± 0.7	0.2	0.6	26 ± 40	< 0.001
Left ventricular ejection fraction (%)	46	55.5 ± 3	57.5 ± 4	1.9 ± 6	0.2	3.7	4 ± 11	0.03
Right atrial area (cm2)	45	25.1 ± 7	21.4 ± 8	−3.6 ± 6	−5.6	−1.7	−12 ± 25	< 0.001
Right atrial volume (mL)	45	90.4 ± 41	74.2 ± 48	−16.2 ± 37	−27	−5	−12 ± 42	0.005
RV end-diastolic basal dimension (cm)	48	5.3 ± 0.8	4.6 ± 0.8	−0.5 ± 0.7	−0.7	− 0.3	−8.5 ± 14	< 0.001
RV end-diastolic mid cavity dimension (cm)	48	4.2 ± 0.7	3.9 ± 0.9	−0.3 ± 0.7	−0.5	−0.1	−6.8 ±17	0.005
RV end-diastolic longitudinal dimension (cm)	48	8 ± 1	8 ± 1	−0.04	−0.3	0.2	0.1 ±10	0.75
RV end-diastolic area (cm2)	44	33.9 ± 9	31 ± 8	−2.9	−5	−0.8	−7 ± 22	0.008
TAPSE (cm)	44	1.5 ± 0.5	1.9 ± 0.5	0.42	0.3	0.6	36 ± 43	<0.001
RV function	47
-normal		0 (0)	4 (9)					<0.001
-mild		5 (11)	13 (28)
-moderate		19 (40)	19 (40)
-severe		23 (49)	11 (23)
Leftward shifting of the IVS	45
-Absent		6 (13)	20 (44)					0.001
-Present		39 (87)	25 (56)
Inferior vena cava collapse	40
-Absent		20 (50)	11 (28)					0.035
-Present		20 (50)	29 (72)

Abbreviations: IVS: interventricular septum, RV: right ventricular, TAPSE: tricuspid annular plane systolic excursion.

Table 3.

Doppler echocardiographic determinations before and after a year of parenteral prostacyclin analogue treatment in patients with PAH.

	n	Initial echocardiogram Mean ± SD, n (%)	Echocardiogram at 1 year of treatment Mean ± SD, n (%)	Difference between studies Mean ± SD	Difference between studies		Percentage change Mean ± SD	p (paired student t- or McNemar test)
					Lower 95% CI	Upper 95% CI
Tricuspid regurgitation severity	46
-Mild		9 (20)	26 (57)					<0.001
-Moderate		22 (48)	14 (30)
-Severe		15 (32)	6 (13)
Peak tricuspid regurgitant velocity (m/s)	46	4.3 ± 0.5	3.8 ± 0.6	−0.5 ± 0.6	−0.6	−0.3	−10 ± 14	< 0.001
RVSP (mm Hg)	46	82.4 ± 16	67.4 ± 18	−15 ± 17	−20.1	−9.8	−17 ± 21	< 0.001
RV outflow tract flow acceleration time (ms)	35	55.1 ± 17	73 ± 20	18.4 ± 17	10.7	26.1	43 ± 50	< 0.001
RV outflow tract time-velocity integral (cm)	42	12.6 ± 3.6	17.5 ± 6.6	4.9 ± 7	2.8	7.1	48 ± 66	<0.001
Doppler RV outflow tract notching	40
- Absent		15 (35)	29 (67)					< 0.001
- Present		28 (65)	14 (33)
Ratio of tricuspid regurgitant velocity/RV outflow tract time-velocity integral	40	0.37 ± 0.1	0.25 ± 0.1	− 0.13 ±0.1	−0.2	0.09	−30 ± 28	< 0.001
PVR estimation [18]	38	9 ± 3.2	5.3 ± 2.6	− 3.7	−4.8	−2.7	−38 ± 28	<0.001
Peak E velocity (cm/s)	45	60.5 ± 25	76.8 ± 22	16.2 ± 30	7.1	25	46 ± 76	0.001
E/A ratio	43	0.91 ± 0.5	1.1 ± 0.4	0.19 ± 0.6	0.01	0.38	47 ± 77	0.04
Peak S velocity (cm/s)	33	49 ± 12	54 ± 13	5 ± 10	1.5	8	12 ± 22	0.006
Peak D velocity (cm/s)	33	39 ± 15	48 ± 13	9.1 ± 16	3.4	14.7	33 ± 42	0.002
S/D ratio	33	1.4 ± 0.4	1.2 ± 0.4	−0.18	−0.31	−0.06	−10 ± 24	0.004
Left ventricular diastolic function	47
-Normal		15 (32)	28 (60)					0.02
-Grade I		31 (66)	17 (36)
-Grade II		1 (2)	2 (4)

Abbreviations: PVR: pulmonary vascular resistance, RV: right ventricular, RVSP: right ventricular systolic pressure.

4- Outcomes after receiving parenteral prostacyclin analogues for 1 year

Patients were followed for a median (IQR) of 52.5 (20.5 – 80) months. During this period 18 patients (37.5 %) died, 15 (83 %) from progression of PAH and 3 (17 %) from other conditions (sepsis in 2 patients and gastrointestinal bleeding in 1). The cumulative incidence of death from any cause is shown in figure 2. All but 5 of the 48 patients received prostacyclin analogue treatments until death, transplantation or end of the study. The reasons for discontinuation of the parenteral prostacyclin analogue treatment were recurrent central venous catheter infection (n=2), severe side effects (n=2) and patient decision (n=1). These 5 patients were started on oral and/or inhaled medications and all but one were alive at the end of the study. In the patients that continued treatment with prostacyclin analogues (epoprostenol (n=24, 56 %) and treprostinil (n=19, 44%)), the median (IQR) doses of these medications were 50.5 (24.6–78.5) ng/kg/min for epoprostenol and 98 (66–129) ng/kg/min for treprostinil.

Figure 2. Cumulative Kaplan Meier estimates from the time of the echocardiography after 1-year of parenteral prostacyclin analogues to the time to death from any cause.

5- Echocardiographic changes at 1-year and impact on survival

Overall mortality, adjusted by age and gender, was associated with percentage change in RV mid-cavity dimension (HR per 10% decrease: 0.76 (95% CI: 0.58–0.99)), RV end-diastolic area (HR per 10% decrease: 0.73 (95% CI: 0.57–0.93)), RV outflow tract velocity-time integral (HR per 10% increase: 0.90 (95% CI: 0.83–0.98)), tricuspid valve regurgitation velocity (HR per 10 % decrease: 0.58 (95% CI: 0.37–0.89)), estimated right ventricular systolic pressure (HR per 10 % decrease: 0.79 (95% CI: 0.63–1)) and difference in qualitative RV function (HR per 1 unit of improvement [e.g. moderate to mild]: 0.55 (95% CI: 0.31–0.96)). Other echocardiographic determinations did not reach statistical significance (e-table 2). Results were similar when considering a composite event that included overall death and lung transplantation (data not shown).

Three variables remained significant predictors of long-term mortality in a multivariate model that included age, gender and percentage change in RV mid-cavity dimension, difference in qualitative RV function, RV outflow tract velocity-time integral and tricuspid valve regurgitation velocity. These variables were RV mid-cavity dimension (HR per 10% decrease: 0.68 (95% CI: 0.49–0.93)), RV outflow tract velocity-time integral (HR per 10% increase: 0.87 (95% CI: 0.79–0.96)) and tricuspid valve regurgitation velocity (HR per 10% decrease: 0.53 (95% CI: 0.27–1)). Of these 3 variables, only percentage change in tricuspid valve regurgitation velocity (HR per 10% decrease: 0.61 (95%CI: 0.37–0.98)) remained a significant predictor of long-term mortality in a multivariate model that included baseline NYHA functional class, pulmonary vascular resistance and six-minute walk distance.

The ROC areas under the curve (AUC) were significant for percentage change in RV mid-cavity dimension (AUC=0.69 (95% CI: 0.54–0.82, p=0.04)) and tricuspid valve regurgitation velocity (AUC=0.73 (95% CI: 0.58–0.88, p=0.008)) in predicting mortality after a year of prostacyclin analogue treatment. In the event that the tricuspid valve regurgitation velocity does not decrease during the first year of treatment the sensitivity and specificity for dying is 50% and 86 %, respectively. If the tricuspid valve regurgitation velocity decreases less than 4.5 % during the first year of treatment, the sensitivity and specificity for overall death is 72 and 75 %, respectively (3).

When the 3 patients that died of conditions other than PAH progression were excluded, we observed that change in RV end-diastolic area (HR per 10 % decrease: 0.81 (95% CI: 0.66 – 0.98), pulmonary artery acceleration time (HR per 10 % increase: 0.83 (0.69–1)), estimated right ventricular systolic pressure (HR per 10 mmHg decrease: 0.87 (95 % CI: 0.76–0.98)), tricuspid valve regurgitation velocity (HR per 10 cm/s decrease: 0.68 (95% CI: 0.51–0.92)) and improvement in the leftward shifting of the IVS (HR: 0.38 (95% CI:0.16–0.89) were predictors of mortality due to PAH progression. Using a forward stepwise Cox regression only the change in RV end-diastolic area and tricuspid regurgitation velocity were included in the model.

6- Added value of percentage difference versus baseline or 1-year echocardiographic variables

We compared the change in echocardiographic variables with baseline and 1-year determinations, to assess whether the change augments the prognostic information of each independent component of the equation. Baseline RV basal (HR per 1 cm increase: 2.23 (95%CI: 1.23–4.06)) and RV mid-cavity (HR per 1 cm increase: 2.34 (95%CI: 1.27–4.35)) dimensions were the only baseline echocardiographic variables associated with long-term mortality. RV end-diastolic area (HR per 10 cm² decrease: 0.54 (95% CI: 0.30–0.96)), RV basal (HR per 1 cm increase: 2.79 (95%CI: 1.47–5.30)) and RV mid-cavity (HR per 1 cm increase: 2.67 (95%CI: 1.56–4.58)) dimensions, tricuspid valve regurgitation velocity (HR per 1 m/s decrease: 0.3 (95% CI: 0.12–0.89)) and qualitative RV function (HR per unit [e.g. mild=1 and moderate=2]: 0.46 (95% CI: 0.24–0.87)) at 1-year were significant predictors of survival. The basal and mid-cavity dimension at the year of treatment and the change in RV end-diastolic area and tricuspid valve regurgitation velocity were selected by backward stepwise Cox regression models that included the measure at a year and the difference between echocardiograms.

Discussion

Using detailed assessments, our study demonstrates that a year of parenteral prostacyclin analogue treatment improves several echocardiographic parameters in patients with PAH, suggesting a favorable effect on disease progression in patients that continue treatment for this long. More importantly, we show that the degree of improvement in RV size, peak tricuspid regurgitant velocity, estimated RVSP, RV outflow tract velocity-time integral and subjective RV function are associated with overall survival.

A limited number of studies have investigated the predictive role of echocardiographic parameters in patients with PAH. The presence of pericardial effusion, a reflection of right ventricular dysfunction, has been frequently associated with mortality [20–22]. Other echocardiographic determinations that have demonstrated prognostic implications includes TAPSE [23, 24], right atrial area [20, 25], right ventricular diameter [26] and degree of tricuspid valve regurgitation [24, 25].

Echocardiography permits a serial noninvasive evaluation of cardiac morphology and function; therefore it remains the imaging option of choice to follow patients with PAH. A trial that randomized PAH patients to bosentan or placebo (BREATHE-1) showed a decrease in pericardial effusion, increase in LV diastolic area and improvement in LV systolic eccentricity index, RV to LV diastolic areas ratio, RV ejection time, LV stroke volume and early diastolic filling after 16 weeks of therapy [27]. The effects of long-term epoprostenol therapy on echocardiographic parameters have not been adequately investigated. A study that randomized patients to epoprostenol or placebo demonstrated a beneficial effect of epoprostenol in reducing right ventricular size, curvature of the interventricular septum and maximal tricuspid regurgitant jet velocity at 12 weeks [28]. Long-term studies of epoprostenol treatment failed to show an improvement in RV size but one demonstrated an improvement in RV function [29, 30].

After a year of parenteral prostacyclin analogue treatment, we observed an increase in the left atrial and ventricular sizes, reduction in right atrial and ventricular sizes, improvement in TAPSE, RV outflow acceleration time, RV outflow tract time-velocity integral, qualitative RV function, left ventricular systolic and diastolic function and reduction in tricuspid regurgitation severity, peak tricuspid regurgitant velocity and right ventricular systolic pressure. These favorable changes support that parenteral analogue therapies are certainly efficacious in PAH. However, it is unclear whether the measured changes in the echocardiographic variables have inherent prognostic significance and if this percentage difference has better predictive value than either the baseline or 1-year determinations.

In our cohort, with the exception of RV basal and mid-cavity dimensions, only the echocardiographic values obtained after a year of receiving parenteral prostacyclin analogue treatment and their respective change from baseline were significant predictors of long-term mortality. Particularly, the percentage change in RV end-diastolic area and tricuspid regurgitation velocity were predictors of overall and PAH-associated mortality. Right ventricular failure was the main cause of death in our cohort and we found an improvement in subjective measurements of RV morphology and function after a year of parenteral prostacyclin therapy, suggesting that the RV is able to adapt and remains a critical factor to determine long-term patient’s survival [8].

We have previously reported a high prevalence of grade I LV diastolic dysfunction in patients with severe PAH [31]. Impaired relaxation (i.e. grade I LV diastolic dysfunction) is predominantly due to displacement of the interventricular septum towards the LV during early diastole resulting in a decreased LV chamber size and filling [32–34]. This reciprocal relation of the ventricles (ventricular interdependence) is also observed at the auricular level, given the limited space for the right atrium to expand [35]. After a year of treatment with IV prostacyclin analogues we observed that the LV diastolic function improved and that the transmitral (E/A ratio) increased, in association with an increase in the LA atrium area, LV end-systolic and end-diastolic diameters and decrease in the proportion of patients with leftward displacement of the interventricular septum. Similarly, other investigators found an increase in the E/A ratio after treatment with PAH-specific therapies [27, 36].

There are limitations including the retrospective nature of the present study and that our results only apply to PAH patients who received at least a year of parenteral prostacyclin analogue treatment. Only a few patients had data on RV function parameters that are less afterload-dependent such as tricuspid lateral annular systolic velocity, isovolumic contraction peak velocity or isovolumic acceleration by Doppler tissue and RV longitudinal systolic strain by speckle-tracking imaging [37, 38]. In patients with atrial fibrillation (n=3) we only averaged 3 beats instead of more reliable approaches of averaging ≥ 5 beats or using the index-beat method (a single beat is selected when the ratio of preceding R-R interval (RR1) to pre-preceding R-R interval (RR2) is closer to 1) [39]. Nevertheless, this study adds important information to the sparse evidence regarding the long-term effects of PAH-specific therapy on diverse echocardiographic parameters. Even when there was a marked response to therapy in our cohort, only a few echocardiographic parameters such as changes in tricuspid regurgitant jet velocity, RV morphology and function, predicted long-term survival. Those patients who do not show such favorable echocardiographic changes may need to be treated more aggressively or referred earlier for lung transplant evaluation. Future studies are needed to determine the best time interval to repeat studies (i.e. six-minute walk test, echocardiography, right heart catheterization) after the initiation of PAH-specific therapies and which parameter variation (alone or in combination) is able to predict long-term outcomes.

Conclusions

Echocardiographic parameters that estimate right ventricular systolic pressure and assess RV morphology and function improve after a year of prostacyclin analogue treatment and the degree of change has prognostic implications.

Figure 3. ROC curve analysis of percentage change in tricuspid regurgitation velocity and survival status.

The area under the curve is 0.73 (95% CI: 0.58–0.88, p=0.008). The sensitivity and specificity for dying is 50% and 86 %, respectively, when the tricuspid valve regurgitation velocity does not decrease during the first year of treatment. The sensitivity and specificity for overall death is 72 and 75 %, respectively, when the tricuspid valve regurgitation velocity decreases less than 4.5 % during the same interval.

Abbreviations

HR

hazard ratio

IQR

interquartile range

PAH

pulmonary arterial hypertension

RV

right ventricular

TAPSE

Tricuspid Annular Plane Systolic Excursion