Summary
Background
According to recent USA guidelines, right ventricular (RV) dysfunction can be diagnosed on the basis of a single parameter, such as tricuspid lateral annular systolic velocity (S’) < 10 cm/s or RV fractional area change (RVFAC) < 35%.
Aims
To assess these recommendations in a large unselected cohort of patients awaiting cardiac surgery and evaluate less validated RV function criteria.
Methods
Among the consecutive patients, 413 were prospectively enrolled and underwent comprehensive echocardiography, including S’, RVFAC and other RV parameters (right myocardial performance index; acceleration time, isovolumic velocity and isovolumic acceleration [IVA]; RV dP/dt; isovolumic relaxation time; two-dimensional [2D] strain). We defined subgroups of highly probable RV dysfunction (S’ < 10 cm/s and RVFAC < 35%) and highly probable normal RV function (S’ ≥ 10 cm/s and RVFAC ≥ 35%) as reference groups. Indices of preload and afterload were also recorded.
Results
Of 413 patients, 320 (77.5%) had normal RV function. In 93 patients, S’ and/or RVFAC were abnormal; both were abnormal in 39 (42%) patients. Using our reference groups, IVA ≤ 1.8 m/s 2 and basal 2D strain ≥ –17% were of most value in diagnosing RV dysfunction. IVA was least load dependent while basal 2D strain appeared to be afterload and preload dependent.
Conclusion
In this large population, S’ and RVFAC were sometimes discrepant, supporting the need for a multiparametric approach when evaluating RV function. Among seven less validated criteria, IVA and 2D strain had the best diagnostic value. Unlike 2D strain, IVA was not influenced by loading conditions.
Résumé
Contexte
Les récentes recommandations américaines proposent de diagnostiquer la dysfonction ventriculaire droite (VD) sur au moins un critère validé tel que le pic de vitesse systolique annulaire tricuspide (S’) < 10 cm/s ou la fraction de raccourcissement de surface VD (FRSVD) < 35 %.
Objectifs
Le but de notre étude a été d’évaluer ces recommandations sur une large population de patients non sélectionnés en attente de chirurgie cardiaque, et d’évaluer d’autres indices de fonction VD moins validés.
Méthodes
Parmi les patients, 413 ont bénéficié d’un examen échocardiographique complet incluant S’, FRSVD et d’autres paramètres VD non routiniers : indice de Tei, dP/dt ; temps de relaxation isovolumique ; temps d’accélération, vitesse de la contraction isovolumique et accélération de la contraction isovolumique (IVA) ; 2D strain. Nous avons défini un sous-groupe de patients ayant une forte probabilité de fonction VD normale (S’ ≥ 10 cm/s et FRSVD ≥ 35 %) et un sous-groupe ayant une forte probabilité de dysfonction VD (S’ < 10 cm/s et FRSVD < 35 %) comme nos groupes de référence. Les indices de pré- et post-charge ont également été recueillis.
Résultats
Parmi les 413 patients, 320 (77,5 %) patients ont une fonction VD très probablement normale (S’ et FRSVD normaux). Chez 93 patients, S’ et/ou FRSVD sont anormaux mais seuls 39 (42 %) patients ont les 2 critères simultanément pathologiques. En utilisant nos groupes de référence, l’IVA ≤ 1,8 m/s 2 et le 2D strain basal ≥ –17 % ont la meilleure valeur diagnostique pour détecter une dysfonction VD. De plus, l’IVA est indépendante des conditions de charge alors que le 2D strain est pré- et post-charge dépendant.
Conclusion
Dans cette large population, S’ et FRSVD sont parfois discordants soulignant la nécessité d’une approche multiparamétrique pour évaluer la fonction VD. Parmi 7 nouveaux paramètres, l’IVA et le 2D strain basal ont la meilleure valeur diagnostique. Contrairement au 2D strain, l’IVA reste indépendante des conditions de charge.
Background
Right ventricular (RV) systolic dysfunction has long been recognized as being of prognostic value in various pathological conditions. However, the definition of RV dysfunction using echocardiography has evolved significantly in recent years. In the past, RV function was assessed visually, until studies focusing on tricuspid annular displacement demonstrated convincing results . Since then, numerous echocardiography-Doppler criteria have been proposed and clinical guidelines were published recently by the American Society of Echocardiography (ASE) . The guidelines recommend performing and reporting at least one of the following: RV fractional area change (RVFAC); Doppler-derived tricuspid lateral annular systolic velocity (S’); tricuspid annular plane systolic excursion (TAPSE); and right myocardial performance index (RMPI). Among these criteria, abnormal RV function should be suspected when S’ is < 10 cm/s, TAPSE is < 16 mm, RVFAC is < 35% or RMPI (tissue Doppler) is > 0.55. As proposed by the ASE guidelines, ‘combining more than one measure of RV function may more reliably distinguish normal from abnormal function’. Various combinations are possible, such as S’ and RMPI ; however, on the basis of the quality of validation , the combination of RVFAC with S’ is probably the best.
Interestingly, criteria evaluating RV function based on various physiological approaches are available, such as isovolumic acceleration (IVA; derived from peak isovolumic velocity [IVV] and acceleration time [AT]), RV dP/dt or two-dimensional (2D) peak longitudinal strain. However, normality thresholds are not yet available or are not fully validated. In addition, RMPI using tissue Doppler has been validated less.
The aim of the present study was: to describe the distribution of well-validated criteria (S’, RVFAC) in routine clinical practice; to evaluate the diagnostic value of less validated echocardiographic parameters (RMPI using tissue Doppler, IVA, RV dP/dt and 2D peak longitudinal strain); and to study the influence of loading conditions, as measured by echocardiography, on each parameter.
Methods
Study population
The study group consisted of 422 patients recruited in the Department of Cardiology and Cardiac Surgery at the Clinique Saint-Augustin (Bordeaux, France). This prospective study enrolled patients awaiting cardiac surgery over a 5-month period.
Nine patients (2.1%) were excluded from all subsequent analyses because of inability to record RVFAC or S’; accordingly, data from 413 patients were analysed.
Using recently published ASE guideline thresholds , patients were split into three groups. Group 1 included those with highly probable normal RV function defined by S’ ≥ 10 cm/s and RVFAC ≥ 35%. Group 2 involved those with probable RV dysfunction defined by the presence of one of two criteria: S’ < 10 cm/s or RVFAC < 35%. Group 3 included patients with highly probable RV dysfunction defined by the presence of two criteria: S’ < 10 cm/s and RVFAC < 35%. To test other RV function parameters, groups 1 and 3 only were used, due to the high certainty of RV function assessment in these groups (normal or abnormal).
The local ethics committee (CPP-SOO3, University of Bordeaux 2, France) approved the study protocol and all included patients gave their consent.
Echocardiographic studies
Echocardiographic studies were performed on a Vivid7 or a VividE9 ultrasound system (GE Vingmed Ultrasound SA, Horten, Norway) equipped with multifrequency transducers (1.5–4 MHz). All examinations were performed by experienced sonographers, according to ASE guidelines the day before the surgical procedure, and were stored on a digital workstation for subsequent off-line analysis (EchoPAC PC, GE Vingmed Ultrasound SA, Horten, Norway).
Echocardiographic measurements of right ventricular function
A modified apical four-chamber view focused on the right ventricle (RV) was used to measure RV area, by tracing the RV endocardium both in systole and diastole . RVFAC (%) was defined as ([end-diastolic RV area – end-systolic RV area] / end-diastolic RV area) × 100 .
All measurements using pulsed-wave tissue Doppler imaging (TDI) ( Fig. 1 ) were performed in a four-chamber view focused on the RV using a tissue Doppler mode with a pulsed-wave Doppler sample volume placed in the basal segment of the RV free wall. Special care was taken to ensure optimal image orientation, to avoid underestimation of velocities. The following data were obtained: S’ was defined as the peak longitudinal velocity of the basal RV free wall ; isovolumic relaxation time (IVRT) was defined as the time from the end of the ejection period (S’) to the beginning of tricuspid e’ wave; RMPI obtained by the TDI method (pulsed-wave velocity of the tricuspid annulus) was defined as the ratio of the sum of IVRT and isovolumic contraction time divided by ejection time (S’ duration) ; IVA (m/s 2 ) was defined as the IVV divided by the time interval from onset of the isovolumic wave to its peak velocity (acceleration time [AT]) .
RV dP/dt (mmHg/s) was calculated by measuring the time required for the tricuspid regurgitation (TR) continuous-wave Doppler signal to increase in velocity from 0.5 to 2 m/s . Using the simplified Bernoulli equation, dP/dt was calculated as 15 mmHg divided by this time.
An apical four-chamber view used for 2D strain analysis was obtained using second-harmonic imaging, with frequency, depth and sector width adjusted for frame-rate optimization (60–100/s). All parameters were averaged over three consecutive beats. In postprocessing analysis, the region of interest was obtained by tracing the RV endocardial borders at the level of the basal septum, the apex and the basal free wall at end-systole. However, from this tracing, we limited our analysis to the basal segment for two main reasons: strain measurement reproducibility of the basal segment is better than that of the apical segment , related to a poorer quality image; regional afterload heterogeneity leads to conflicting data between basal and apical 2D strain . The longitudinal myocardial deformation (defined as the peak longitudinal systolic strain) was expressed as a percentage of the longitudinal shortening of basal segment in systole compared with in diastole.
Echocardiographic measurements of right ventricular load
RV preload assessment included end-diastolic RV area and inferior vena cava (IVC) maximal and minimal diameters obtained from a subcostal view at end-expiration, after sniff test and quiet breathing. The IVC collapse index (%) was defined as ([maximal IVC – minimal IVC] / maximal IVC) × 100. Maximal IVC diameter and collapse index were used to assess preload according to the ASE guidelines . Normal right atrial (RA) pressure (3 mmHg) was defined by an IVC diameter < 2.1 cm and a collapse > 50%, whereas an IVC diameter > 2.1 cm that collapsed by < 50% suggested high RA pressure (15 mmHg). RA area was measured at end-systole (four-chamber view).
RV afterload assessment included the following:
- •
systolic pulmonary artery pressure (SPAP) was determined from peak TR velocity using the simplified Bernoulli equation and combined with an estimated RA pressure (based on the IVC collapse index);
- •
pulmonary vascular resistance (PVR) was estimated using peak TR velocity and RV outflow tract (RVOT) time velocity integral (TVI) – the formula published by Abbas et al. was applied, i.e. PVR = ([peak TR velocity / RVOT TVI] × 10) + 0.16;
- •
pulmonary AT derived from pulsed-wave Doppler recordings was defined as the time interval from onset of the RVOT anterograde flow to its peak velocity.
Reproducibility
Twenty-five data sets were chosen randomly to assess intraobserver variability by repeating measurements 4 weeks apart (J. Peyrou). Interobserver agreement was assessed by a second observer (M. Simon) using the same data sets.
Statistical analyses
Continuous data are expressed as means ± standard deviations and discrete parameters as absolute numbers and percentages. The groups were tested using a Chi 2 test to compare categorical parameters. The clinical and echocardiographical data from the patients’ group were compared using the two-sample Student’s t test or Wilcoxon’s rank sum (Mann-Whitney) non-parametric test, as appropriate, according to the variance R test. The relationships between additional parameters (RMPI, dP/dt, IVRT, IVV, AT, IVA, basal 2D strain) and the validated criteria of RV function (S’ and RVFAC) as well as the load dependency of parameters were tested by means of Pearson’s correlation. Receiver operating characteristic curve analysis was performed to test the diagnostic accuracy for discrimination between patients with RV dysfunction (group of patients with high probability of RV dysfunction) and those with normal RV function, and to determine optimal cut-off values. Intraobserver and interobserver reproducibility were analysed using the Pearson correlation coefficient, the concordance correlation coefficient according to the Lin method and the coefficient of variability. All P values were two-sided and values < 0.05 were considered statistically significant. All statistical analyses were performed using Stata software version 11.0 (StataCorp LP, College Station, TX, USA).
Results
Patient characteristics
Four-hundred and thirteen patients (mean age 70.3 ± 10.3 years) were enrolled, of whom 63% were awaiting valve surgery, 49% coronary artery bypass grafting and 3% other cardiac surgery (myxoma resection, atrial septal defect closure, pericardiectomy, aortic dissection repair, coarctation repair). Thirty-four patients (8%) had chronic atrial fibrillation. Table 1 shows patients’ characteristics in each group.
| Variable | Group 1 (normal RV function) ( n = 320) | Group 2 (probable RV dysfunction) ( n = 54) | Group 3 (highly probable RV dysfunction) ( n = 39) |
|---|---|---|---|
| Sex-ratio | 2.42 | 2.8 | 3.2 |
| Age (years) | 70 ± 10.3 | 72 ± 9.1 | 71.5 ± 10.7 |
| Cardiovascular risk factors | |||
| Hypertension | 215 (67.2) | 30 (55.5) | 22 (56.4) |
| Dyslipidaemia | 205 (64.1) | 33 (61.1) | 19 (48.7) |
| Overweight (BMI > 25 kg/m 2 ) | 193 (60.3) | 30 (55.5) | 24 (61.5) |
| Tobacco use | 105 (32.8) | 16 (29.6) | 14 (35.9) |
| Diabetes mellitus | 598 (18.7) | 12 (22.2) | 7 (17.9) |
| Familial history | 23 (7.1) | 5 (9.3) | 3 (7.7) |
| Treatment | |||
| ACEIs/ARBs | 173 (54.1) | 29 (53.7) | 21 (53.8) |
| Diuretics | 115 (35.9) | 26 (48.1) a | 24 (61.5) a |
| Beta-blockers | 135 (42.2) | 28 (51.8) | 22 (56.4) |
| Calcium channel blockers | 80 (25.0) | 6 (11.1) a | 6 (15.4) |
| Statins | 217 (67.8) | 34 (63.0) | 23 (59.0) |
| Platelet inhibitors | 188 (58.7) | 28 (51.8) | 22 (56.4) |
| Coronary artery disease | 168 (52.5) | 22 (40.7) | 22 (56.4) |
| Severe valvular heart disease | |||
| Aortic stenosis | 142 (44.4) | 23 (42.6) | 18 (46.1) |
| Aortic regurgitation | 12 (3.7) | 1 (1.9) | 0 (0) |
| Mitral stenosis | 11 (3.4) | 2 (3.7) | 1 (2.5) |
| Mitral regurgitation | 25 (7.8) | 15 (27.8) a | 3 (7.7) |
| Main surgical procedures | |||
| CABG | 168 (52.5) | 22 (40.7) | 14 (35.9) |
| Valve replacement | 182 (56.9) | 32 (59.2) | 17 (43.6) |
| Valve repair | 23 (7.2) | 12 (22.2) a | 2 (5.1) |
| AAA repair | 36 (11.2) | 5 (9.2) | 2 (5.1) |
a Only significant P values (< 0.05) are indicated, comparing RV dysfunction groups (group 2 or 3) with normal subjects (group 1).
Age, sex, blood pressure and cardiovascular risk factors did not differ significantly between normal and RV dysfunction groups. Compared with normal subjects (group 1), patients with probable or highly probable RV dysfunction (groups 2 and 3) had a lower left ventricular ejection fraction (65.2 ± 11.3%, 59.9 ± 13.9% and 50.8 ± 17.3%, respectively; P < 0.001) and a higher SPAP (36.6 ± 12.2%, 45.9 ± 19.2% and 53.8 ± 21.9%, respectively; P < 0.001).
Right ventricular function parameters
Of 413 patients, 320 (77.5%) had normal RV function (group 1: S’ ≥ 10 cm/s and RVFAC ≥ 35%) and 93 patients (22.5%) had suspected RV dysfunction (S’ < 10 cm/s and/or RVFAC < 35%). Among these 93 patients, 54 (58%) had a discrepancy between indices, suggesting probable RV dysfunction (group 2: S’ < 10 cm/s but RVFAC ≥ 35% or S’ ≥ 10 cm/s but RVFAC < 35%) and 39 patients had concordant indices suggesting highly probable RV dysfunction (group 3: S’ < 10 cm/s and RVFAC < 35%).
As shown in Table 2 , RMPI, IVRT, IVV, IVA and basal 2D strain were statistically different between patients with normal RV function and patients with probable (group 2) or highly probable (group 3) RV dysfunction. dP/dt and AT were not statistically different.
