Article Text
PDF
Abstract
Background Right ventricular (RV) dysfunction is associated with adverse outcomes in patients with pulmonary hypertension (PH). This systematic review and meta-analysis evaluated the prognostic value of RV free-wall longitudinal strain (RVfwLS), compared with other RV parameters in PH.
Methods We searched for articles presenting the HR of two-dimensional RVfwLS in PH. HRs were standardised using the within-study SD. The ratio of HRs of a 1 SD change in RVfwLS versus systolic pulmonary arterial pressure (SPAP), systolic tricuspid annular velocities (s’-TV), RV fractional area change (FAC) or tricuspid annular plane systolic excursion (TAPSE) was calculated for each study, after which we conducted a random model meta-analysis. Subgroup analysis regarding the type of outcome, aetiology of PH and software vendor was also performed.
Results Twenty articles totalling 2790 subjects were included. The pooled HR of a 1 SD decrease of RVfwLS was 1.80 (95% CI: 1.62 to 2.00, p<0.001), and there was a significant association with all-cause death (ACD) and composite endpoints (CEs). The ratio of HR analysis revealed that RVfwLS has a significant, strong association with ACD and CE per 1 SD change, compared with corresponding values of SPAP, s’-TV, RVFAC or TAPSE. RVfwLS was a significant prognostic factor regardless of the aetiology of PH. However, significant superiority of RVfwLS versus other parameters was not observed in group 1 PH.
Conclusions The prognostic value of RVfwLS in patients with PH was confirmed, and RVfwLS is better than other RV parameters and SPAP. Further accumulation of evidence is needed to perform a detailed subgroup analysis for each type of PH.
Trial registration number UMIN Clinical Trials Registry (UMIN000052679).
- Pulmonary Arterial Hypertension
- Ultrasonography
- Meta-Analysis
Data availability statement
Data are available in a public, open access repository.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Right ventricular (RV) dysfunction is strongly associated with adverse outcomes in patients with pulmonary hypertension (PH), regardless of WHO classification
Right ventricular free-wall longitudinal strain (RVfwLS) has high reproducibility, and is one of the most useful echocardiographic parameters for predicting prognosis.
WHAT THIS STUDY ADDS
This systematic review and meta-analysis found that RVfwLS has better prognostic value than other echocardiographic RV parameters and systolic pulmonary arterial pressure in patients with PH.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
The confirmed high prognostic value of RVfwLS in patients with PH provides a strong rationale for prioritising RVfwLS measurement in busy, daily clinical practice.
Introduction
Pulmonary hypertension (PH) is a grave diagnosis, with a 5-year survival rate of 55%.1 Regardless of WHO classification, right ventricular (RV) dysfunction is strongly associated with death.2–5 The RV is more sensitive to afterload than the left ventricle (LV), and RV remodelling is often irreversible, making right heart failure a direct cause of death.6 7 Therefore, early detection of RV impairment facilitates timely and appropriate therapy, thereby improving survival.
RV ejection fraction (RVEF) by cardiac MRI or three-dimensional echocardiography (3DE) is the reference standard for evaluating RV function, but its clinical applicability remains relatively low.8–12 Analysis of RV function with transthoracic two-dimensional echocardiography (2DE) is the routine method of choice. For example, systolic lateral tricuspid valve annular velocities (s’-TV), RV fractional area change (FAC) and tricuspid annular plane systolic excursion (TAPSE) are commonly used to assess RV systolic function.13 The prognostic value of FAC and TAPSE in patients with PH has been reported.2 14 However, those three parameters have significant drawbacks for evaluating RV function due to the complexity of RV geometry and function.15
RV free-wall longitudinal strain (RVfwLS) with 2DE speckle-tracking analysis is highly reproducible, angle independent and correlates best with 3DE RVEF in patients with PH.16 Based on these advantages of RVfwLS, two previous meta-analyses have demonstrated the prognostic usefulness of RVfwLS in patients with PH.17 18 However, neither of these publications examined whether RVfwLS is more closely associated with adverse outcomes than other RV function and haemodynamic parameters. Accordingly, we conducted a systematic review and performed a meta-analysis to examine the prognostic value of RVfwLS in patients with PH, and we compared it directly with those of other RV functional and haemodynamic parameters.
Material and methods
Search strategy and selection of articles
This systematic review and meta-analysis was registered in the UMIN Clinical Trials Registry (registration no. UMIN000052679) and was conducted following PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analysis) guidelines.19 The PICO approach was used to formulate a search question as follows: P (population): patients with PH regardless of aetiology; I (intervention or exposure): participants experienced diagnosis of PH by echocardiography or catheterisation; C (comparison): comparison of echocardiographic parameters in patients with and without clinical outcomes; O (outcome): reported one or more of the following outcomes: all-cause death or composite of cardiopulmonary outcomes. Two authors searched published literature from three electronic databases (PubMed, Embase and Scopus) using the keywords “speckle tracking”, “right ventricular”, “strain” and “prognosis” on 29 November 2022 and 18 February 2023. Details of the search strategy are shown in online supplemental table 1. One publication was added by reviewing reference lists of included studies. After the initial search and exclusion of duplicates, two authors (YN and TK) independently screened study titles and abstracts. After the screening, a full-text assessment was performed of the remaining studies in accordance with inclusion criteria. Inclusion criteria were as follows: (1) full-text papers in peer-reviewed journals, (2) studies reporting the mean and SD of RVfwLS by 2DE speckle-tracking analysis, (3) studies that reported outcomes of all-cause mortality or a composite endpoint, including cardiopulmonary events and death, and (4) studies reporting the HR of RVfwLS using univariable or multivariable Cox proportional hazard analysis. If there were several publications from the same group of authors in which the study populations overlapped, we selected the article with the largest number of study subjects. All publications were limited to adult subjects in the English language.
Supplemental material
Data extraction
Two authors (YN and MT) extracted data regarding first author name, year of publication, journal information, number of subjects, mean age, gender, type of PH, primary endpoint, follow-up duration, number of events, mean±SD of RVfwLS and HR with its 95% CI. RVfwLS was presented as an absolute value. HR with a 95% CI and mean±SD of systolic pulmonary arterial pressure (SPAP), s’-TV, RVFAC and TAPSE were also extracted, if available.
Statistical analyses
HRs and 95% CIs were rescaled using the SD for each parameter to compare studies and variables.20 21 This rescaling allows a direct comparison of the HR for a 1 SD decrease in RVfwLS, s’-TV, FAC and TAPSE and HR for a 1 SD increase in SPAP. Then differences in log-transformed HRs of RVfwLS versus SPAP, s’-TV, FAC or TAPSE were calculated in each study. These estimates were pooled using a random-effect model (DerSimonian-Laird), and finally converted to an HR scale. A ratio of HR >1 means that a 1 SD decrease in RVfwLS is related to a larger hazard increase than a 1 SD decrease (or increase) in other RV parameters for an association with outcomes. Subgroup analysis was performed to assess the influence of the type of outcome, WHO classification and software vendor on the results. Publication bias was assessed using a funnel plot and Egger’s test. If publication bias was observed, trim-and-fill analysis was performed. I2 was used to assess between-study heterogeneity. Find-outlier analysis was performed when between-study heterogeneity was suspected. Meta-regression analysis was performed to determine effects of publication year, sample size, follow-up duration, age and sex on the HR of RVfwLS. Study quality assessment was performed using the Downs and Black Quality Assessment Checklist.22 Statistical analysis was performed using R V.4.2.2 (The R Foundation for Statistical Computing, Vienna, Austria).
Results
A PRISMA flow chart is presented in online supplemental figure 1. Among 4048 articles retrieved from three databases, 20 articles and 21 patient groups with 2790 subjects were finally included in this meta-analysis.23–42 Tables 1 and 2 present characteristics of included studies and patients, respectively. The pooled mean value of RVfwLS was 17.0±2.5%, reflecting impaired RV systolic function.43 Five studies included only WHO group 1 PH, and one study included only group 2 PH. The remaining studies involved patients with various types of PH. The primary endpoint was all-cause mortality in 9 studies, and 11 other studies used a composite endpoint. Follow-up duration ranged from 13 to 77.7 months. Of the 20 articles, SPAP, s’-TV, FAC and TAPSE were reported in 10, 4, 12 and 14 studies, respectively. Seventeen studies reported results of univariable analysis, and three studies reported results of multivariable analysis after adjusting several covariates. Reasons for full-text exclusion are summarised in online supplemental table 2. Online supplemental table 3 provides a summary of the study quality analysis.
Characteristics of the study population (n=2790 from 20 studies)
Characteristics of included studies
Outcomes
Prognostic value of RVfwLS
Figure 1 depicts forest plots for HR of RVfwLS. A 1 SD decrease in RVfwLS was significantly associated with outcome, with an HR of 1.80 (95% CI: 1.62 to 2.00, p<0.001; heterogeneity: I2=40%, p=0.032). According to the I2 value, between-study heterogeneity was likely. Find-outlier analysis revealed that Hasselberg et al’s study was an outlier.39 After that study was removed, a 1 SD decrease in RVfwLS was still associated with outcome, and heterogeneity diminished (HR: 1.83, 95% CI: 1.66 to 2.03, p<0.001; heterogeneity: I2=29%, p=0.107). Online supplemental figure 2 depicts the funnel plot, and Egger’s test indicated asymmetry (p=0.008). Trim-and-fill analysis added eight studies to make funnel plot symmetric; however, the prognostic significance of RVfwLS was maintained (HR: 1.64, 95% CI: 1.46 to 1.84, p<0.001; heterogeneity: I2=47%, p=0.003).
Forest plots of the HR of right ventricular free-wall longitudinal strain for association with outcome. Observed and estimated HRs and their 95% CIs are shown in each study. Pooled estimates are shown by random models. HK, Hoffman and Kringle; seTE, standard error of treatment effect; TE, treatment effect.
Comparison between RVfwLS and conventional RV parameters
RVFwLS versus SPAP
In 10 studies reporting both HR of RVfwLS and HR of SPAP in the same cohort, the HR per 1 SD change for RVfwLS was 1.23 times greater (95% CI 1.10 to 1.38, p=0.003) than the corresponding value of SPAP for the association of the outcome with no heterogeneity (I2=0%, p=0.649) (figure 2A).
Forest plots of the ratio of HRs per 1 SD reduction of right ventricular free-wall longitudinal strain (RVfwLS) and other right ventricular parameters: (A) RVfwLS versus systolic pulmonary arterial pressure (PAP); (B) RVfwLS versus systolic tricuspid annular velocity; (C) RVfwLS versus right ventricular fractional area change (FAC); (D) RVfwLS versus tricuspid annular plane systolic excursion (TAPSE). The x-axis represents the log-transformed ratio of HR. HK, Hoffman and Kringle; seTE, standard error of treatment effect; TE, treatment effect.
RVfwLS versus s’-TV
In four studies reporting both HR of RVfwLS and of s’-TV, random model meta-analysis revealed that HR of a 1 SD decrease in RVfwLS was 1.42 times larger (95% CI 1.04 to 1.95, p=0.038) than the HR of a 1 SD decrease in s’-TV for the association of adverse outcome with no heterogeneity (I2=0%, p=0.469) (figure 2B).
RVfwLS versus FAC
In 12 studies in which both HR of RVfwLS and HR of FAC were described, the HR of a 1 SD decrease in RVfwLS was 1.22 times greater (95% CI 1.07 to 1.39, p=0.008) than the HR of a 1 SD decrease in FAC with no heterogeneity (I2=0%, p=0.808) (figure 2C).
RVfwLS versus TAPSE
In 14 studies in which both HR of RVfwLS and HR of TAPSE were provided, the HR of a 1 SD decrease in RVfwLS was 1.21 times greater (95% CI 1.03 to 1.42, p=0.022) than the HR of a 1 SD decrease in FAC with low heterogeneity (I2=27%, p=0.164) (figure 2D).
Subgroup analysis
Type of outcome
Nine studies employed all-cause death (ACD), and 12 studies employed a composite endpoint as a primary outcome. Regardless of the type of outcome, RVfwLS was significantly associated with future prognosis (online supplemental figure 3).
Aetiology of PH
Five studies and two studies included only patients with WHO group 1 and group 2 PH, respectively. The remaining 14 studies included various types of PH. RVfwLS was a significant prognostic factor for all types of PH (online supplemental figure 4). However, when restricted to studies of WHO group 1, there was no significant difference in prognostic utility between RVfwLS and SPAP, FAC or TAPSE (online supplemental figure 5).
Software vendor
Numbers of studies that used GE, TomTec, Epsilon and Syngo were 10, 3, 3 and 2, respectively. Studies using GE and Epsilon showed statistical significance in the prognostic utility of RVfwLS (online supplemental table 6). Strain measurements exhibit vendor dependency; however, it is difficult to draw conclusions from these results because of large differences in numbers of studies.
Meta-regression analysis
Meta-regression analysis was performed to examine the effect of publication year, sample size, follow-up duration, age and sex on the HR of RVfwLS. Only follow-up duration was associated with the HR of RVfwLS (estimate: −0.005, p=0.034). Therefore, the HR of RVfwLS decreased significantly with longer follow-up.
Discussion
To the best of our knowledge, this is the first meta-analysis to compare the prognostic value of RVfwLS and conventional RV function and haemodynamic parameters in patients with PH. Major findings are as follows: (1) RVfwLS is strongly associated with adverse outcomes in patients with PH, even after accounting for heterogeneity and publication bias; (2) RVfwLS shows a significantly stronger correlation with adverse outcomes than a 1 SD change in other RV functional and haemodynamic parameters; and (3) subgroup analyses revealed that RVfwLS is associated with prognosis regardless of WHO classification.
Previous studies
RV decreases contraction and output under increased afterload more readily than LV.6 Thus, RV systolic dysfunction is observed regardless of the type of PH and may regulate cardiac output and prognosis. European Society of Cardiology/European Respiratory Society guidelines for diagnosis and treatment of PH published in 2022 revised the risk classification methodology.5 The prognostic impact of right ventricular contractility was emphasised, and descriptions of RVEF by cardiac MRI and TAPSE/SPAP were added. However, difficulty remains in regard to accessibility and reproducibility for cardiac MRI and TAPSE/SPAP measurements, respectively. On the other hand, RVfwLS is highly reproducible and easy to determine.16 Although the RV exhibits complex morphology and dynamics, the free-wall contributes the most to contraction, with more muscle mass in subendocardial, longitudinally oriented fibres than in epicardial transverse fibres.44 45 Therefore, RVfwLS analysis in apical four-chamber view demonstrates the validity of assessing RV motion.
As far as we know, only two systematic reviews and meta-analyses on the prognostic value of RVfwLS in patients with PH have been published previously.17 18 First, Shukla et al published a systematic review and meta-analysis to assess the prognostic value of RVfwLS in 1001 patients with PH from 10 studies.17 All-cause mortality was registered in six articles, and there were 193 deaths (18%). RVfwLS was significantly associated with death (HR: 1.14, 95% CI: 1.11 to 1.8, p<0.001, per 1% change). However, other RV parameters, such as TAPSE and FAC, were not assessed as continuous variables. Second, Hulshof et al18 included not only studies assessing RVfwLS, but also studies assessing RV global longitudinal strain in 1169 patients with PH from 11 studies. Seven studies describing a composite endpoint comprised 821 patients. Five studies presented all-cause mortality as an endpoint, including 399 patients. RV longitudinal strains (RVLS) were significantly associated with both all-cause mortality and composite endpoints. However, values of HR were presented as the relative reduction of RVLS, and it was quite difficult to rescale as a 1% change.
Current study
We found that several clinical questions remained after considering results of the above studies. First, the RV functional and haemodynamic indices most associated with prognosis were not discussed. Second, whether the prognostic value of RVfwLS is associated with the aetiology of PH has not been examined. Third, it is unclear whether advances in treatment and changes in guidelines have changed the importance of RVfwLS.
Since the last systematic review and meta-analysis conducted in 2019, nine papers have been published assessing the prognostic value of RVfwLS in patients with PH. A 1 SD reduction in RVfwLS represented a 1.80-fold (95% CI: 1.62 to 2.00, p<0.001) increase in the risk of an adverse outcome. These results are consistent with previous meta-analyses.17 18 Prognostic usefulness of RVfwLS was directly compared with other conventional RV parameters and SPAP. HR of a 1 SD reduction in RVfwLS was significantly greater than the HR per 1 SD change in all four comparisons. This may be due to angle independency, low interobserver variability and reduced dependency on loading conditions of strain analysis.46
According to results of subgroup analysis, RVfwLS is significantly associated with future prognosis, regardless of the aetiology of PH. However, there was no statistically significant difference between RVfwLS and other RV parameters in patients with WHO group 1 PH. This may be due to the small sample size (n=491), and further accumulation of evidence is needed. RV geometry and systolic function may differ between pre-capillary and post-capillary PH due to differences in LV end-diastolic pressure and systolic function. This difference could influence the association between RVfwLS and FAC; however, to date, no studies included this kind of assessment. Multidirectional assessment of RV motion using software developed by Lakatos et al may offer some insights.47
Meta-regression analysis revealed that the prognostic utility of RVfwLS was not affected by publication year. Although a number of PH drugs have been developed in recent years, the importance of RV systolic function assessment by RVfwLS in predicting prognosis remains unchanged. This may be because most current PH drugs reduce afterload, but do not improve RV function itself.48 However, RV dysfunction is not just a prognostic factor, but is now considered a therapeutic target, and translational research is ongoing.49 In this new era of PH treatment, repeatable and reproducible RV function indices are becoming increasingly important, and we believe RVfwLS may become as common as LV global longitudinal strain in patients with PH.
However, it is unclear whether RVfwLS should be routinely measured in all patients. Recently, we published a study assessing the prognostic value of RVfwLS in patients with aortic stenosis.50 In that meta-analysis, RVfwLS was not significantly superior to other RV parameters in predicting adverse outcomes. Further study is required to determine in what diseases RVfwLS should be measured.
Limitations
This study was influenced by the variability of the original studies because of the nature of meta-analysis. Second, studies included in this meta-analysis were not uniform in population characteristics, follow-up duration or endpoints. This heterogeneity may affect HRs. However, meta-regression analysis indicated that only follow-up duration was associated with the HR of RVfwLS. Third, we did not analyse the prognostic utility of RV–pulmonary artery (PA) coupling in this study. TAPSE/SPAP is called RV–PA coupling and is thought to reflect the efficiency of RV stroke work.51 Some studies reported the prognostic utility of RV–PA coupling in patients with PH.52 However, the number of studies is insufficient, and the physiological meaning is not well understood. It is not equal to Ees/Ea because both TAPSE and SPAP are loading dependent. Further studies are required. Fourth, we included only parameters of systolic function in this study, but not tricuspid regurgitation, right atrial area, inferior vena cava diameter and eccentricity index because these deviate from the purpose of our study. Fifth, we used studies that employed both ACD and composite endpoints as outcomes and calculated HRs for mixed outcomes. Because ACD and the composite endpoint are clinically very different outcomes, separate analyses were also performed for each endpoint. Sixth, the possibility of interobserver variability of RVfwLS, SPAP, FAC, TAPSE and s’-TV across included studies should be taken into consideration. Seventh, we could not assess the incremental value of combining different echocardiographic parameters for predicting adverse outcomes because the number of studies analysing such methods was insufficient. Eighth, RVfwLS is routinely measured only in a limited number of laboratories. One possible reason for this may be the analysis time: Li et al reported that manual measurement required 103 s for analysis, whereas fully automated analysis required only 8 s.53 As fully automated analysis becomes more widespread, RVfwLS measurements may become more common. Ninth, results differed with the software vendor, but it is difficult to explain why. Il’Giovine et al reported significant differences between RVfwLS by GE and Philips and RVfwLS by TomTec, but the differences were small, and the causes of these differences were not clear.54 Prospective studies on the relationship between intervendor variability of RVfwLS and prognostic utility are required. Finally, the number of included studies for direct comparison was not large enough to perform more detailed subgroup analysis, considering the type of PH and software used for the analysis. In particular, the result that RVfwLS did not show a prognostic advantage over other RV functional parameters in WHO group 1 was unexpected. However, we are not prepared to conclude that RVfwLS is less useful in patients with WHO group 1 PH. All case analyses, including at least 985 WHO group 1 patients, showed the utility of RVfwLS, and only two or three papers each on WHO group 1 have directly compared RVfwLS with other RV function indices. Further studies directly comparing RVfwLS and other RV parameters are required.
Conclusions
RVfwLS is associated with adverse outcomes in patients with PH. Moreover, superior prognostic values of RVfwLS over SPAP, s’-TV, RVFAC and TAPSE were observed in the ratio of HR analysis. However, sufficient evidence is lacking, and further studies are needed to draw conclusions regarding each type of PH. We believe RVfwLS will facilitate better management of patients with PH.
Data availability statement
Data are available in a public, open access repository.
Ethics statements
Patient consent for publication
Ethics approval
Not applicable.
Acknowledgments
We thank Dr Victor Chien-Chia Wu for his assistance with the literature search.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Contributors YN—data curation, analysis and writing. TK—data curation. KN—supervision. MT—conceptualisation, methodology, analysis, review and editing, guarantor.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.