Tricuspid annular plane systolic excursion is preserved in young patients with pulmonary hypertension except when associated with repaired congenital heart disease

Abstract

Aims

Tricuspid annular plane systolic excursion (TAPSE) is a measure of right ventricular (RV) longitudinal function that correlates with functional status and mortality in adults with pulmonary hypertension (PH). The diagnostic and predictive value of TAPSE in children with PH has not been fully examined. We aimed to define TAPSE across aetiologies of paediatric PH and assess the correlation between TAPSE and measures of disease severity.

Methods and results

TAPSE measurements were obtained in 84 children and young adults undergoing treatment for PH and 315 healthy children to establish z-scores at moderate altitude for comparison. The relationships between TAPSE and echocardiographic, biomarker, and functional measures of disease severity between aetiologies were assessed. TAPSE z-scores in PH patients with congenital heart disease (CHD) repaired with open cardiac surgery (n = 20, mean −2.73) were significantly decreased compared with normal children and patients with other aetiologies of PH (P < 0.001) but did not reflect poorer clinical status. TAPSE z-scores in children with idiopathic PH (n = 29, −0.41), unrepaired CHD (n = 11, −0.1), and PH related to systemic disease (n = 14, −0.39) were not different from normal. TAPSE correlated modestly with brain natriuretic peptide, echocardiographic function parameters, and functional class except in patients with repaired CHD.

Conclusion

Children with PH maintain normal TAPSE values early except when associated with repaired CHD. Superior RV adaptation to high afterload in children compared with adults may account for this finding. Reduced TAPSE after repair of CHD does not correlate with functional status and may reflect post-operative changes rather than poor function primarily due to PH.

Introduction

Pulmonary hypertension (PH) is a heterogeneous, progressive disease of elevated pulmonary arterial pressure and pulmonary vascular resistance that results in right heart failure and, ultimately, death. PH can be idiopathic or associated with multiple underlying conditions. While overall survival in children with PH has improved, mortality remains higher in some aetiological groups, including those with repaired congenital heart disease (CHD).^1–3 The reference standard for diagnosing and monitoring PH is right heart catheterization. In clinical practice, non-invasive measures of disease severity including echocardiography, 6-min walk distance (6MWD), and biochemical markers of heart failure such as brain natriuretic peptide (BNP) and NT-proBNP are useful in predicting survival and monitoring response to therapy.^4–6

Echocardiographic findings associated with PH have been well described.⁷ However, assessment of right ventricular (RV) function is difficult due to its complex geometry, anterior thoracic location, and trabeculations that limit demarcation of the endocardium. Despite these limitations, standard two-dimensional, Doppler and tissue Doppler (TDI) echocardiographic modalities are useful in estimating pulmonary artery pressure (PAP) and RV function in PH.^8–10 Tricuspid annular plane systolic excursion (TAPSE) is an easy, reproducible measure of RV longitudinal function that correlates with cardiac output and predicts mortality in adult patients with PH.¹¹ Normal TAPSE values including z-scores based on patient age have been established in paediatric patients, and TAPSE has been shown to be associated with outcome in idiopathic PH.^12,13 The effect of moderate altitude on normal right ventricular longitudinal function in children has not been determined. We studied TAPSE in a heterogeneous population of children and young adults with PH to assess the clinical utility of TAPSE in evaluating disease severity in this population. The specific aims of our study were as follows:

1. To establish normal TAPSE values and z-scores based on BSA for children at moderate altitude

2. To determine if TAPSE in children with PH of different aetiologies differs from normal values

3. To determine if TAPSE correlates with other clinical markers of disease severity in paediatric PH

Methods

Patient population

All patients <25 years of age (1–25 years) diagnosed with PH at age <18 years and enrolled in an Internal Review Board (IRB)-approved protocol entitled Prospective Evaluation of Adolescents and Children with Pulmonary Hypertension (PEACH) were reviewed for inclusion in this retrospective study (n = 91). Of these, 7 patients were excluded from the study (4 with absent TAPSE measurement, 2 with epicardial pacemakers, 1 primary cardiomyopathy), leaving 84 patients for review. Enrolled patients were previously diagnosed with PH based on catheterization measurement of mean PAP >25 mmHg (n = 81). The remaining 3 patients had systolic PAP greater than half the systemic systolic blood pressure and an indexed pulmonary vascular resistance (PVRi) >3 Wood Units × m² but did not have a mean PAP of >25 mmHg. Average systolic PAP, mean PAP, and PVRi at diagnosis were 68 ± 22 mmHg, 47 ± 18 mmHg, and 10.2 ± 6.5 Wood Units × m², respectively. All patients were receiving therapy for PH at the time of the reviewed echocardiogram. Informed consent was obtained from participants included in the study.

An additional 315 healthy patients (age 5 days to 19 years) referred to outpatient cardiology clinic for evaluation of murmur, family history of CHD or chest pain who had technically adequate TAPSE, and heart disease excluded by echocardiography were reviewed. The University of Colorado IRB approved the retrospective use of normal echocardiograms for research purposes.

Echocardiographic measurements

All echocardiograms were performed using a standard protocol including two-dimensional, spectral and colour-flow Doppler, TDI, and M-mode examinations on GE Vivid 7 (GE Healthcare, Milwaukee, WI, USA) or Phillips IE-33 (Phillips Medical Systems, Andover, MA, USA) ultrasound systems. Images were recorded digitally and analysed by an imaging specialist at our institution to produce a clinical report that was used to obtain qualitative assessment of RV function, degree of septal flattening, and estimated RV pressure. Septal flattening was defined as moderate if the septum formed a ‘D’ shape of the LV, mild if there was less flattening and severe if the septum bowed into the LV during systole. RV systolic pressure was calculated from the maximum velocity of the tricuspid regurgitant jet using the modified Bernoulli equation. A right atrial pressure of 5 mmHg was assumed for all patients. Peak systolic velocity of the right ventricular free wall (RV S′) was measured using TDI with a 5 mm sample volume placed at the base of the RV lateral wall from the apical four-chamber view. RV fractional area change (FAC) was obtained as described previously.¹⁴ TAPSE was obtained from two-dimensional-guided M-mode recordings from the apical four-chamber view with the cursor placed at the free wall of the tricuspid annulus.¹⁵ TAPSE and RV S′ measurements from one to three consecutive cardiac cycles as available were analysed by a single investigator (A.H.) and results averaged.

Clinical variables

BNP and NT-proBNP measurements and 6MWD were performed at the discretion of patients’ clinical provider. 6MWD was performed in patients >8 years of age who did not have physical or cognitive limitations to completing the testing. Values were included in the analysis if they were performed within 2 weeks of the measured TAPSE without interval change in clinical status or therapy. The 6MWD was performed according to a standardized protocol on a flat, 30-m course. 6MWD was converted to percent-predicted distance walked using reference equations for children and young adults.^16,17 Echocardiograms and 6MWD were performed on the same day.

BNP samples were assayed on an I-STAT system using the two-site enzyme-linked immunosorbant assay (Abbott Laboratories, Chicago, IL, USA). NT-proBNP samples were measured on an electrochemiluminescnce immunoassay (Mayo medical laboratories, Toche Diagnostics, Indianapolis, IN, USA).

Modified World Health Organization (WHO) functional class was assigned to patients by clinical providers and retrieved from the electronic medical record.

Statistical analysis

All statistics were performed using R 2.14.1 and SAS 9.3. Continuous variables are expressed as mean ± standard deviation (SD). To assess the inter- and intra-reader reliabilities for TAPSE measurements, Bland–Altman plots and intra-class correlation coefficients (ICC) were evaluated for a sample of 30 normal patients.

Due to potential differences in our patient population based on different racial diversity, obesity prevalence, and evaluation at moderate elevation, reference values (z-scores) for normal TAPSE were established based on BSA. A non-parametric spline model was used. The mean TAPSE trajectory of $E(\text{TAPSE})=f(\text{BSA})$ was approximated by $\mathrm{\Sigma }{\beta }_j{\beta }_j(BSA)$, a linear basis expansion using cubic B-splines $[{\beta }_j(BSA)]$ (R Package SPLINES). The z-score for each subject was then calculated using equation $z_i=[\text{TAPS}{\text{E}}_i-f(\text{BS}{\text{A}}_i)]/RMSE$, where RMSE is the root mean square error of the model. Based on the fitted model, TAPSE z-scores for PH patients were calculated and compared with those of normal children using a two-sample t-test. One-way analysis of variance (ANOVA) followed by Tukey multiple comparisons or Kruskal–Wallis tests were performed as appropriate to study differences among the aetiology groups within the PH patients.

The relationships between TAPSE and other markers of disease severity were assessed using one-way ANOVA for categorical variables and Pearson's $(\gamma )$ or Spearman's correlation $(\rho )$ for continuous variables.

Results

The demographic and clinical characteristics of patients with PH as a whole and by aetiology are shown in Tables 1 and 2, respectively. There were no significant differences in age or BSA between groups. Estimated RV pressures by echocardiography were highest in patients with unrepaired CHD but were not significantly different between other groups. 6MWD was significantly lower in the unrepaired CHD group and the ‘other’ group compared with both idiopathic pulmonary arterial hypertension and repaired CHD groups.

Table 1.

Characteristics of patients with PH (n = 84)

Parameter
M/F	42/42
Age (years)	11.1 ± 6.4
BSA (m2)	1.13 ± 0.44
Aetiology of PH
Idiopathic	29 (35%)
Congenital heart disease	41 (49%)
Repaired	30
Isolated PDA	2
ASD ± other L → R shunt	9
VSD ± PDA	2
AVSD	5
Coarctation + shunt lesion	6
Transposition of the great arteries	2
Other	4
Unrepaired	11
Isolated PDA	1
ASD ± other L → R shunt	4
VSD ± PDA	3
AVSD	1
Other	2
Other	14 (17%)
Lung disease	8
Autoimmune	3
Hematological	2
Genetic disorder (3q deletion)	1
RVSP/SBP	0.58 ± 0.22
WHO functional class
I	31 (37)
II	31 (37)
III	16 (19)
IV	6 (7)

ASD, atrial septal defect; AVSD, atrioventricular septal defect; BSA, body surface area; PDA, patent ductus arteriosus; RVSP/SBP, ratio of right ventricular systolic pressure to systemic systolic blood pressure by echocardiography; VSD, ventricular septal defect; WHO, World Health Organization.

Table 2.

Characteristics of patients by aetiology of PH

Variable	Idiopathic PH (n = 29)	Repaired CHD (n = 30)	Unrepaired CHD (n = 11)	Other (n = 14)
Demographics
Gender M/F	16/13	17/13	3/8	6/8
Age (years)	11.6 ± 6.6	9.8 ± 6.6	12.7 ± 6.5	12.0 ± 5.6
BSA (m2)	1.2 ± 0.4	1.0 ± 0.5	1.1 ± 0.4	1.2 ± 0.5
Echocardiographic parameters
RVSP/SBP (n)	0.6 ± 0.2 (26)	0.6 ± 0.2 (27)	0.8 ± 0.2 (8)	0.6 ± 0.2 (13)
TR severity: n (%)
Trivial	15 (52)	12 (40)	5 (45)	6 (43)
Trivial–mild	6 (21)	4 (13)	0	2 (14)
Mild	7 (24)	12 (40)	4 (36)	3 (21)
Moderate	1 (3)	2 (7)	2 (18)	3 (21)
Septal flattening: n (%)
None	12 (41)	12 (40)	1 (10)	2 (14)
Mild	9 (31)	13 (43)	3 (30)	4 (29)
Moderate	5 (17)	3 (10)	5 (50)	6 (43)
Severe	3 (10)	2 (7)	2 (20)	2 (14)
Subjective RV
Dysfunction: n (%)
Normal	24 (83)	25 (83)	8 (73)	9 (64)
Low normal–mild	4 (14)	2 (7)	1 (9)	5 (36)
Moderate–severe	1 (3)	3 (10)	2 (18)
RV FAC (n)	0.42 ± 0.13 (26)	0.45 ± 0.13 (28)	0.51 ± 0.10 (11)	0.46 ± 0.09 (14)
RV S′ (n)	12.6 ± 1.6 (26)	9.1 ± 2.8 (26)	11.9 ± 3.1 (10)	12.4 ± 2.5 (14)
Clinical variables
WHO class: n (%)
I	9 (31)	17 (57)	1 (9)	4 (29)
II	17 (59)	8 (27)	3 (27)	3 (27)
III	3 (10)	4 (13)	6 (55)	3 (27)
IV		1 (3)	1 (9)	4 (36)
BNP (n)	58 ± 92 (24)	140 ± 177 (20)	144 ± 234 (8)	42 ± 25 (11)
NT-proBNP (n)	210 ± 331 (24)	402 ± 595 (23)	337 ± 441 (9)	189 ± 138 (11)
6MWD (n)	566 ± 74 (20)	508 ± 122 (11)	468 ± 88 (5)	430 ± 125 (8)
% predicted	(86 ± 11)	(76 ± 18)	(71 ± 14)	(66 ± 19)

BNP, brain natriuretic peptide; PAP, pulmonary artery pressure; PVRi, indexed pulmonary vascular resistance; RV, right ventricle; RVSP/SBP, ratio of right ventricular systolic pressure to systemic systolic blood pressure; WHO, World Health Organization; 6MWD, 6-min walk distance.

TAPSE results in normal children

The relationship between TAPSE and BSA in our normal paediatric population at moderate altitude is shown in Figure 1 (MSE = 0.24, adjusted-R² = 0.60, P < 0.001). There was excellent agreement for both intra-reader (ICC 0.99) and inter-reader (ICC 0.97) TAPSE measurements.

Figure 1.

TAPSE values for BSA in normal population at altitude. Relationship between BSA (m²) and TAPSE (cm) in normal children is shown (adjusted-R² = 0.60). The mean is marked by the solid line. Z-scores of ± 2 are represented by the dashed lines.

TAPSE in paediatric PH

Our BSA-based TAPSE z-scores were significantly lower in patients with PH-repaired CHD compared with normal patients and those with other aetiologies of PH (Figure 2). This reduction was related primarily to repair requiring open cardiac surgery on cardiopulmonary bypass (CPB) (z-score −2.71 ± 1.29) rather than catheter-based intervention or ‘off-pump’ surgical repair via thoracotomy (z-score 0.13 ± 2.02, P < 0.001) (Figure 3). These findings also held true when using available age-based normal values for TAPSE. Non-bypass cases included catheter occlusion of left-to-right shunting lesions (n = 5), surgical PDA ligation (n = 1), and coarctation repair with spontaneous or subsequent catheter occlusion of concomitant shunt lesions (n = 4). TAPSE z-score was not lower in patients requiring ventriculotomy. Time since diagnosis (range 1–23 years) was not related to TAPSE z-score in the surgically repaired group but longer time since diagnosis showed a significant association with TAPSE (r = 0.47, P < 0.01) in the non-CHD aetiologies (Figure 4). Patients undergoing repair using CPB were not significantly different from those who did not with regards to RV systolic pressure, RV FAC, septal flattening, PVRi at the time of diagnosis, or WHO functional class.

Figure 2.

Comparison of TAPSE z-scores by aetiology of PH. TAPSE z-scores are significantly reduced only in patients with PH and repaired CHD. Z-scores for patients with idiopathic PH, PH-unrepaired CHD, and other aetiologies did not differ significantly from the normal population. P-values reflect Tukey multiple comparisons of aetiological groups with the repaired CHD group.

Figure 3.

Comparison of TAPSE z-scores in patients undergoing surgical repair of CHD. Comparison of TAPSE z-score in patients who underwent surgical repair with and without CPB. Mean z-score for those with surgical repairs on bypass was −2.71 ± 1.29. Those repaired off bypass had mean TAPSE z-score of 0.13 ± 2.02, P < 0.001.

Figure 4.

Differences in the relationship between TAPSE Z-score and duration of disease for different aetiologies of PH. TAPSE z-score is reduced but not dependent on duration of disease in patients with repaired CHD. TAPSE is preserved in patients with unrepaired CHD or Eisenmenger syndrome. TAPSE remains normal in most patients up to 15 years after initial diagnosis in patients with other aetiologies of PH, but longer duration of illness is associated with a fall in TAPSE z-score. Dotted lines mark the range of normal z-scores.

TAPSE and clinically utilized markers of disease severity

There was a trend towards lower TAPSE z-scores in patients with echocardiographic measures of more severe PH including qualitatively assessed RV dysfunction (P = 0.059) and worsening septal flattening (P = 0.024) (Figure 5). TAPSE z-score correlated strongly with S′ of the right ventricular free wall (r = 0.78). There was a small but significant correlation between TAPSE and FAC (r = 0.23, P = 0.04) that improved when patients with repaired CHD on CPB were excluded (r = 0.36, P < 0.01). There was no significant correlation between TAPSE z-score and RV systolic pressure by echocardiogram.

Figure 5.

Relationship between TAPSE z-scores and qualitative assessment of right ventricular effects of PH by echocardiogram. Box plots illustrating the decreased TAPSE z-scores in patients with (A) RV dysfunction compared with those with qualitatively normal RV systolic function (P = 0.059) and (B) septal flattening compared with normal septal configuration (P = 0.027).

Among all patients with PH, there was a small but significant correlation between TAPSE z-score and BNP (ρ = −0.29, P = 0.019). In patients with PH-repaired CHD, NT-proBNP but not BNP correlated with TAPSE z-score (ρ = −0.50, P = 0.031). TAPSE z-score did not significantly correlate with 6MWD in the heterogeneous population with PH or in those with PH-repaired CHD. TAPSE z-score fell with worse functional class except in those who underwent CPB (Figure 6).

Figure 6.

TAPSE Z-score by WHO class in patients with and without exposure to bypass. TAPSE z-score is similar in patients with PH exposed to cardiopulmonary bypass regardless of functional class. TAPSE is significantly lower in patients with worse functional status in those without bypass exposure (I vs. II, P = 0.07, I vs. III/IV P = 0.01, II vs. III/IV ns). *P = 0.01.

Discussion

In this study, TAPSE values for BSA in normal children were established at moderate altitude at our institution. While previously published normal TAPSE values were reported across a range of BSA, to our knowledge, equations to calculate z-score based on BSA have not been published. The creation of our BSA- rather than age-based z-scores provided a more valid comparison for our patient population consisting of small-for-age patients with chronic disease. Our range of normal values followed a similar relationship to BSA as previous paediatric data; however, normal values were lower for older children in our population. The reason for this difference is unclear but may be related to differences in population. Alternatively, RV longitudinal function may be reduced in the setting of moderate altitude as has been demonstrated previously at high altitude.¹⁸ In children with PH, we found that TAPSE is significantly reduced only in children with repaired CHD utilizing CPB. There was a trend towards lower TAPSE with longer duration of disease; however, most patients maintained TAPSE z-scores in the normal range. Open cardiac surgery requiring CPB may be an important factor contributing to RV longitudinal function in children with PH. These findings have not been previously reported in this population, and whether this effect is related to CPB or other surgical factors is yet undetermined. Whatever the cause, the altered RV mechanics following repair of CHD may become clinically important in the failing RV and contribute to the poorest outcomes that have been described in this group.^1–3,19

This is the first study to evaluate the relationship between TAPSE and markers of disease severity in children with PH. We found associations between TAPSE and other commonly used echocardiographic markers of right ventricular performance in paediatric PH including septal flattening, subjective RV function, FAC, and RV S′, which has previously been shown to correlate with TAPSE in normal patients.²⁰ TAPSE did not correlate with estimated RV pressure by echocardiogram, signifying that RV pressure overload alone does not result in decreased longitudinal function. Non-echocardiographic measures of clinical severity such as WHO functional class was associated with TAPSE z-score in those without CHD; however, even those with severe limitations did not have consistently reduced TAPSE. Interestingly, serum biomarkers of heart failure correlated modestly with TAPSE, especially in the repaired CHD group. TAPSE did not correlate with 6MWD, perhaps due to limitations of the submaximal nature of this test in children with PH.²¹

Our results emphasize that differences exist between adult and paediatric PH. In adults with PH, TAPSE has been shown to be an effective marker of RV function and predictor of outcome, correlating with haemodynamic measures, exercise capacity, and NYHA class.^11,22 Importantly, TAPSE <1.8 cm is a strong predictor of mortality in adults with PH.^8,11 Compared with adult studies evaluating TAPSE, our patients have similar baseline haemodynamic measurements, especially when taking into account their lower systemic blood pressures.^9,11 However, our cohort was considerably less symptomatic than those in adult studies. It is known that children with PH have preserved cardiac output and later onset of right heart failure with higher relative PAP and resistance when compared with adults.²³ TAPSE represents regional basal longitudinal function, which may be preserved earlier in the disease process despite elevated PAPs. Differences in adaptation by the paediatric right ventricle to increased afterload as well as earlier stages of disease may explain our finding of preserved TAPSE and poorer correlation with clinical measures of disease in children with PH compared with adults.

Studies investigating tricuspid annular motion specifically in children with PH have shown reduced tricuspid annular displacement in patients with PH and both repaired CHD and structurally normal hearts when compared with age-matched infants and children.^13,24–26 TAPSE has been predictive of outcomes in paediatric patients with idiopathic PH but not PH-repaired CHD in studies by Kassem et al. and Okumura et al. These results contrast with our findings and may reflect a difference in disease severity, management, or patient population.^13,27 While our patients had moderate-to-severe PH at the time of diagnosis, the measured PAP and PVRi in this study were significantly lower than those reported in others. The paediatric right ventricle may continue to preserve longitudinal function until end-stage PH. The trend towards lower TAPSE in patients with longer duration of disease suggests that TAPSE can be followed as a measure of PH in patients without CHD.

Reduction in TAPSE and tricuspid S′ in children with PH-repaired CHD has been shown previously by Koestenberger et al. who demonstrated values below the expected value for age by 5 and 10 years, respectively, after surgical repair.²⁸ Our results differ in that TAPSE z-scores in this study were consistently low regardless of the length of time since CPB. The reason for this difference is unclear; however, reduced RV longitudinal function is a well-documented though not fully understood phenomenon following cardiac surgery in adults.^29,30 Despite reduction in longitudinal function, RV FAC and global RVEF are preserved, suggesting that global function may not be well reflected by TAPSE post-operatively.³¹ Reduced TAPSE persists after full recovery from surgery.³² This same alteration in longitudinal function while maintaining preserved global RV function likely occurs following repair of CHD. Koestenberger et al. demonstrated that TAPSE is significantly decreased following repair of tetralogy of Fallot, but correlation of TAPSE with cardiac MRI-derived RVEF is variable.^33–35 In patients with hypoplastic left heart syndrome, TAPSE is significantly reduced following the Norwood procedure despite preserved RV FAC.³⁶ These data are consistent with our findings of reduced TAPSE in the absence of other echocardiographic or functional measures of disease severity in paediatric PH with repaired CHD. Cardiac surgery thus results in reduced tricuspid annular motion without obligate reduction in RV performance in hearts exposed to normal PAPs. It is unclear whether exposure of the RV to elevated afterload further reduces TAPSE beyond the effects of surgery alone, and this should be investigated further.

Basal longitudinal contraction clearly plays an important role in global RV function in both the normal heart and in PH; however, apical function and transverse shortening are also important components of RV function in PH.^27,37–39 This may be especially true in patients with PH-repaired CHD in whom longitudinal function is altered. RV FAC or transverse contraction may better capture global function in this group. There is ongoing need to better understand the mechanics of the RV in adults and children with PH.

Limitations

This study was performed retrospectively with patients already on therapy, many with controlled disease. Catheterization data was available for all patients only at the time of diagnosis; therefore, haemodynamics could not be correlated with echocardiographic parameters at the time of TAPSE measurement. Additionally, this is not a longitudinal study, and each patient was examined at only one point in their disease course. Changes in TAPSE in individual patients over time may demonstrate differences in TAPSE with disease progression. We were unable to investigate all aetiologies of PH separately due to limitations of sample size. The ‘other’ group includes PH of multiple aetiologies that may not result in the same effects on the right ventricle. However, the population of patients studied in this analysis is representative of the population of children in PH referral centres.^1–3,19

Conclusions

TAPSE is significantly reduced in paediatric patients with PH associated with CHD repaired with open cardiac surgery, and is generally preserved in children with PH of other aetiologies. Preservation of TAPSE in our population may be explained by the later development of RV failure in children than in adults with similar PAPs. Reduced TAPSE after repair of CHD does not correlate with functional status and may reflect post-operative changes rather than poor function primarily due to PH.

Funding

D.D.I. is funded by the National Center for Advancing Translational Sciences National Institutes of Health grant UL1 TR 000154 and by the National Heart Lung and Blood Institute National Institutes of Health grant 5R01HL114753.

Conflict of interest: None declared.