Right ventricular-pulmonary artery coupling for prognostication in acute pulmonary embolism

Abstract

Aims

Acute pulmonary embolism (PE) increases pulmonary pressure and impair right ventricular (RV) function. Echocardiographic investigation can quantify this mismatch as the tricuspid annular plane systolic excursion (TAPSE) to pulmonary arterial systolic pressure (PASP) ratio. The aim of the study was to investigate the prognostic capabilities of TAPSE/PASP ratio in patients with acute PE.

Methods and results

We utilized the Registro Informatizado Enfermedad TromboEmbolica registry to analyse consecutive haemodynamically stable PE patients. We used multi-variable logistic regression analyses to assess the association between the TAPSE/PASP ratio and 30-day all-cause mortality across the strata of European Society of Cardiology (ESC) risk categories. We included 4478 patients, of whom 1326 (30%) had low-risk, 2425 (54%) intermediate-low risk and 727 (16%) intermediate-high risk PE. Thirty-day mortality rates were 0.7%, 2.3% and 3.4%, respectively. Mean TAPSE/PASP ratio was 0.65 ± 0.29 in low-risk patients, 0.46 ± 0.30 in intermediate-low risk and 0.33 ± 0.19 in intermediate-high risk patients. In multi-variable analyses, there was an inverse association between TAPSE/PASP ratio and 30-day mortality (adjusted OR 1.32 [95% CI 1.14–1.52] per 0.1 decrease in TAPSE/PASP). TAPSE/PASP ratio below optimal cut-points was associated with increased mortality in low- (<0.40, aOR: 5.88; 95% CI: 1.63–21.2), intermediate-low (<0.43, aOR: 2.96; 95% CI: 1.54–5.71) and intermediate-high risk patients (<0.34, aOR: 4.37; 95% CI: 1.27–15.0). TAPSE/PASP <0.44 showed net reclassification improvement of 18.2% (95% CI: 0.61–35.8) vs. RV/LV ratio >1, and 27.7% (95% CI: 10.2–45.1) vs. ESC risk strata.

Conclusion

Decreased TAPSE/PASP ratio was associated with increased mortality. The ratio may aid in clinical decision-making, particularly for intermediate-risk patients for whom the discriminatory capability of the current risk stratification tools is limited.

Graphical Abstract

Open in new tabDownload slide

Introduction

In patients with pulmonary embolism (PE), death may result from a mismatch between right ventricular (RV) function and increased pulmonary arterial pressure.^1,2

Transthoracic echocardiography (TTE) is standard tool for assessment of RV function and pulmonary artery pressure and thereby a potential mismatch.³ RV function can reliably be quantified using the tricuspid annular plane systolic excursion (TAPSE), and the pulmonary arterial systolic pressure (PASP) can also be estimated echocardiographically.⁴ Reduced TAPSE or increased PASP are associated with increased mortality.^5,6 TTE is central in risk stratification,^1,3,7 and enhanced methods for identifying patients at risk of deterioration requiring intervention vs. those suitable for discharge are essential.

The TAPSE/PASP ratio describes the echocardiographic RV-to-pulmonary artery (RV-to-PA) coupling. The ratio has been developed and investigated in patients with pulmonary hypertension but its use in acute PE is novel emphasizing the rationale to further explore its capabilities in acute PE (Figure 1). Previous studies found that a reduced TAPSE/PASP ratio was associated with adverse outcomes. However, these studies were limited by restriction to certain PE patients and small study size.^8–10 These studies have been unable to determine if the TAPSE/PASP ratio improved risk prognostication in PE.

Our aim was to examine the TAPSE/PASP ratio in a large PE population using the Registro Informatizado Enfermedad TromboEmbolica (RIETE) registry. We hypothesized that the TAPSE/PASP ratio would be associated with mortality overall, and across the strata of PE severity.

Methods

Study design and population

The RIETE registry is a prospective registry and the world's largest registry on venous thromboembolism (VTE). The design and methodology of the RIETE registry have been previously described.¹¹ Regular audits are conducted to ensure data accuracy. The protocol was approved by the Ethics Committee of Hospital Germans Trias i Pujol (number PR(AG)213/2020) and by all participating centres. All patients provided informed consent.

The current study was an observational study based on data from the RIETE registry from September 2009 (the time when echocardiographic parameters of interest were included in RIETE) until 31 August 2023. We included adult patients with confirmed PE based on CT pulmonary angiogram, high-probability V/Q scan or invasive pulmonary angiogram, and who had TTE performed within 3 days of PE diagnosis. We excluded high-risk PE patients with systemic hypotension (n = 877) and those with missing TAPSE and/or PASP values (n = 15 615).

Outcomes

The primary outcome was 30-day all-cause mortality, with secondary outcomes of 30-day PE-related mortality and 7-days all-cause mortality. Our exploratory tertiary outcome was a composite of 7-days mortality or clinical deterioration defined as hypotension, treatment with vasopressors, intubation and mechanical ventilation, cardiac arrest, treatment with systemic thrombolysis, extracorporeal membrane oxygenation (ECMO) or surgical embolectomy. Several of the variables in the tertiary outcome were not included in the RIETE registry until November 2019, why the analyses of this outcome were restricted to patients included thereafter (n = 992 included).

Echocardiography

The study was based on site-reported RV parameters, why PASP calculation was based on uniform definition since all RIETE sites adhere to the same TTE protocol. For assurance, we recently validated RIETE echocardiographic parameters against core laboratory analysis. TAPSE and PASP exhibited the highest reliability,¹² supporting the use of site-reported information in this study (Figure 1).

In order to risk classify according to ESC guidelines, we captured RV dysfunction on TTE defined as a right-to-left ventricular ratio (RV/LV) > 1, RV basal diastolic diameter > 42 mm, or RV free wall hypokinesis.^1,4 TAPSE and PASP were not included in these definitions since they were variables under investigation in the study.

Risk stratification

According to European Society of Cardiology (ESC) guidelines,¹ we defined low-risk PE if the simplified pulmonary embolism severity index (sPESI) was zero and there was absence of RV dysfunction and elevated troponin levels (normal or missing). Intermediate-low risk PE was defined as sPESI ≥ 1 and either RV dysfunction on TTE or elevated troponin (according to each RIETE center). Intermediate-high risk PE was defined as sPESI ≥ 1 and RV dysfunction on TTE and elevated troponin. In cases where sPESI = 0 but the patient had RV dysfunction on TTE or had elevated troponin, the patient was categorized as intermediate-low risk.¹

Statistical analysis

Data were presented as mean ± SD if normally distributed and median [interquartile range, IQR] if not. We used the Kruskal–Wallis test to compare quantitative variables, and the χ² test to compare qualitative variables. We performed bi- and multi-variable logistic regression analyses adjusting for age, sex, smoker status, past medical history of chronic lung disease and ESC risk stratification group. Odd ratios (OR) and 95% CI were calculated per 0.1 decrease in TAPSE/PASP ratio. To control statistical power, we used hierarchical analyses and investigated if an association between TAPSE/PASP and primary outcome was present. If so, we would proceed to intermediate- and low-risk groups, and if present, we would proceed to intermediate-low and intermediate-high risk groups. We produced receiver operating characteristic curves and established the optimal TAPSE/PASP cut-off points as the point that maximized overall sensitivity and specificity, and OR based on these cut-points were calculated. The net reclassification improvement (NRI) was used to investigate any added value of the TAPSE/PASP to risk stratification compared with RV/LV ratio and ESC risk strata. Patients with TAPSE/PASP values higher vs. lower than the established cut-off point was compared with the Kaplan–Meier method and log-rank-test. A P < 0.05 was considered statistically significant. IBM SPSS Statistics (Version 25) was used for all analyses.

Results

Of 20 970 haemodynamically stable patients with confirmed PE and TTE performed within the first 3 days in RIETE, 4478 (21.4%) had available data on both the TAPSE and PASP and were included in the study. Most of the excluded patients had missing TAPSE or PASP values (Figure 2). Patient characteristics are presented in Table 1. Supplementary material online, Table S1 shows characteristics of excluded patients.

Median time from PE diagnosis to TTE was 1 [IQR 0–2] day, and echocardiographic findings are shown in Table 2. The majority of patients received low-molecular-weight heparin as initial therapy (see Table 3). Ninety-one patients (2%) experienced the primary outcome, fifteen (0.3%) the secondary outcome of PE-related death and thirty-four patients (0.8%) died within 7 days. The tertiary composite outcome occurred to 59/992 (5.9%) patients.

The mean TAPSE/PASP ratio was 0.49 ± 0.31 mm/mmHg. The TAPSE/PASP ratio was negatively associated with the primary outcome with OR 1.35 (95% CI 1.19–1.52) per 0.1 decreases in TAPSE/PASP ratio in bivariate analysis, which remained significant in multi-variate analysis (adjusted OR 1.32 [95% CI 1.14–1.52]). Similar OR point estimates were observed for the secondary and tertiary outcomes (Table 4).

The optimal cut-point for the TAPSE/PASP ratio in our large, unselected PE population to predict 30-days all-cause mortality was 0.44 (Table 2) with an AUC of 0.66 (95% CI 0.61–0.71). For TAPSE/PASP ratio above this cut-point, we found a sensitivity of 80.0% (95% CI 71.7–88.3%), specificity of 49.2% (95% CI 47.7–50.6%), positive predictive value (PPV) of 3.1% (95% CI 2.4–3.8%), and negative predictive value (NPV) of 99.2% (95% CI 98.8–99.6%).

Using the optimal cut-point of 0.44 for the entire population, a lower TAPSE/PASP was associated with increased risk of all-cause mortality (OR 3.87, 95% CI 2.30–6.50), 30-days PE-related mortality (OR 13.3, 95 % CI 1.75–101.2) and 7-days all-cause mortality (OR 9.88, 95% CI 3.01–32.4). Similar ORs were observed in multi-variable analysis (Figure 3 and Table 4). Adjusted ORs were consistent between sex, age > 65 vs. ≤ 65 years or risk category (Figure 3).

Kaplan–Meier analysis of 30-days all-cause mortality showed significant difference between patients with TAPSE/PASP below vs. above 0.44 (log-rank 29.38, P < 0.001), see Figure 4.

TAPSE/PASP <0.44 reclassified patient risk with NRI = 18.2% (CI 95% 0.61–35.8) vs. RV/LV ratio >1. For reclassification into a higher stratum when patients were low- or intermediate-low risk but TAPSE/PASP was below 0.44, or reclassification into a lower stratum when patients were intermediate-low or -high but TAPSE/PASP was ≥ 0.44, the NRI was 27.7% (95% CI 10.2–45.1).

Risk sub-groups

A total of 1326 (30%) patients belonged in the low-risk; 2425 (54%) in the intermediate-low risk and 727 (16%) in the intermediate-high risk category according to ESC criteria. The outcomes and optimal cut-point for each risk-category are shown in Table 3.

Compared with low-risk patients, intermediate risk patients had lower TAPSE and higher PASP. We observed a stepwise decrease in TAPSE/PASP ratio with ESC PE severity, i.e. mean the ratio was 0.65 ± 0.29 for low-risk patients, 0.46 ± 0.30 (P < 0.001 vs. low-risk) for intermediate-low risk patients and 0.33 ± 0.19 for intermediate-high risk patients (P < 0.001 vs. low-risk). See Table 2.

TAPSE/PASP ratio was, consistently, associated with adverse outcome in each risk category of patients in both bi- and multi-variable analyses. Using individual cut-points for each subgroup yielded similar results but more refined ORs (Table 4). Among the patients with intermediate-high risk patients with a TAPSE/PASP ratio < 0.34 mm/mmHg, 30-day mortality was 4.9% (2/41) for patients receiving systemic thrombolysis, 7.7% (1/13) for the mechanical lysis group and 4.5% (18/400) for patients not receiving thrombolysis. For intermediate-high risk patients with TAPSE/PASP ≥ 0.34 mmHg, no patients who received any kind of thrombolysis died (0/22) whereas 30-day mortality was 1.6% (4/251) among patients not receiving thrombolysis.

We calculated sensitivity, specificity and positive and NPVs for each risk category using the representative optimal TAPSE/PASP cut points. For the intermediate-high risk group, we observed only slight improvement in sensitivity to 84.0% (95% CI 69.6–98.4%) and PPV to 4.6% (95% CI 2.7–6.6%). See Supplementary material online, Table S2.

Discussion

The present study is the largest of its kind demonstrating that the echocardiographic TAPSE/PASP ratio is independently associated with adverse outcomes in haemodynamically stable patients with acute PE. For each 0.1 decrease in the TAPSE/PASP ratio, there was 26% increased odds of mortality. The association remained significant across different clinical risk categories. Notably, the TAPSE/PASP ratio, as a measure of RV-to-PA uncoupling, gradually decreases with higher risk categories. The study extended traditional risk stratification including NRI showing significant reclassification.

Uncoupling in acute pulmonary embolism

An acute PE leads to a sudden increase in RV afterload.¹³ The RV responds through dilatation, resulting in reduced RV function. A mismatch between RV function and afterload may culminate in RV failure.¹ This match is recognized as ventriculo-arterial coupling known from invasive pressure-volume loops.¹⁴ Experimental studies utilizing pressure-volume loop recordings have confirmed RV-to-PA uncoupling in acute PE.^15,16

The TAPSE/PASP ratio has been validated as a cost-effective, repeatable and non-invasive echocardiographic measure for assessing RV-to-PA coupling.^17,18 Our study shows that with increasing risk categories, the TAPSE/PASP ratio decreases, suggesting a progressive uncoupling with more severe PE (Graphical abstract).

Clinical implications

Echocardiography, while not diagnostic for PE, plays a pivotal role in risk stratification.³ Traditionally, the focus of echocardiographic risk-stratification has been on RV dilatation or isolated RV function.^5,19 However, the TAPSE/PASP ratio offers an integrative approach to assessing RV function in relation to its specific afterload and has shown superiority to individual measures like TAPSE and PASP.^8,10 The measurements can be easily performed at the bedside^4,20,21 even with POCUS, which increases the utility of the ratio in acute patients requiring immediate risk stratification. Consequently, the TAPSE/PASP ratio may be effortlessly implemented.

The present study revealed an association between the TAPSE/PASP ratio and adverse outcomes, consistent with previous smaller studies.^8–10,22,23 These prior studies focused on intermediate- and high-risk PE patients, whereas our research extends this knowledge to a broader spectrum of PE severity. The association remained significant among patients >65 years old, which is noteworthy as the utility of echocardiography in older PE patients has been a subject of debate.²⁴

The clinical applicability of the TAPSE/PASP ratio extends to chronic pulmonary hypertension (PH).^17,25,26 The optimal TAPSE/PASP cut-point in PH patients is lower (∼0.2). Normal values for TAPSE/PASP, as well as the cut-offs for detecting RV dysfunction, range between 0.7 and 2.0.^27–29 We found that the optimal cut-point for acute PE lies between these values, at ∼0.4, supporting previous studies.^9,10 A recent study on intermediate-high risk PE patients identified an optimal cut-point of 0.33, which is almost identical to our observation in this sub-group.³⁰

Prognostic performance of the TAPSE/PASP ratio in acute pulmonary embolism

Using this cut-point of 0.44, we demonstrated a significant risk reclassification and a very high NPV. While several tools already exist for identifying low-risk patients who can be safely discharged, such as the sPESI score and Hestia criteria,³¹ it is important to note that the 99% NPV also encompasses low-risk patients who nevertheless underwent TTE, as well as intermediate-risk patients typically admitted for observation. Our reporting of sensitivity and specificity may therefore be useful in clinical decision-making.

However, our study revealed a low PPV to predict 30-days all-cause mortality. There are some possible explanations. Most importantly, we observed few events, which from a mathematical standpoint will reduce PPV regardless of a test's specificity. The low PPV in the present study is comparable to studies on Bova score, ESC risk stratification or troponins where PPV is single-digit but higher when prognosing a composite and more frequent outcome than all-cause mortality.^32–34 Zuin et al.³⁰ observed mortality rates twice that of ours and calculated higher sensitivity, specificity and, interestingly, PPV of 32.4% supporting the ratio’s prognostic performance. Second, the physiological rationale of the ratio would argue PE-related death or deterioration as ideal outcomes, and we did see the OR point estimate for 30-days PE related death several fold that of 30-days all-cause mortality. We chose all-cause mortality as primary outcome as it was the endpoint in sPESI development and in ESC risk stratification.^1,35 Other studies have also shown low PPV for different risk tools for prognostication of deterioration.^33,36

Study limitations

Several limitations warrant consideration. First, a substantial proportion (77,7%) of PE patients with TTE were excluded due to missing TAPSE or PASP values, since TAPSE and PASP were not part of RIETE for the first many years of enrolment, and that these values are not mandatory in the RIETE registry. This exclusion mainly affected low-risk patients (see Supplementary material online, Table S1) but may have introduced a selection bias and a risk of error. Second, the interpretation of the tricuspid regurgitation gradient is sometimes used as a proxy for PASP when clinicians fail to add an estimated right atrial pressure. This interpretation may vary depending on local echocardiographic approaches and could affect external generalizability. Finally, the study observed few events, as a considerable number of patients were in lower risk categories. This makes analyses moderately robust.

Conclusion

The TAPSE/PASP ratio is associated with all-cause mortality in a substantial cohort of stable acute PE patients. The optimal cut-off is identified as 0.44. This ratio declines from low-risk to higher-risk PE patients, indicating an increasing degree of RV-to-PA uncoupling.

Supplementary material

Supplementary material is available at European Heart Journal: Acute Cardiovascular Care online.

Funding

National Institutes of Health (R01-HL168040-01 to M.D.L., D.M.D., and C.K.).

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.

References

Author notes