Doppler echocardiography: The noninvasive method for accurately estimating stroke volume in ischaemic cardiogenic shock

Abstract

Accurate stroke volume (SV) assessment is crucial in ischaemic cardiogenic shock. While pulmonary artery catheterisation remains the gold standard, its invasive nature necessitates reliable noninvasive alternatives. However, the literature on echocardiographic SV accuracy is inconsistent. This study evaluated commonly used echocardiographic techniques—Doppler-derived and Simpson’s method—against invasive thermodilution in 39 patients. SV by Doppler showed strong correlation (r = 0.91, p < 0.0001) and minimal bias, whereas the Simpson’s method exhibited weaker correlation and more significant underestimation. These findings suggest Doppler echocardiography as the more accurate noninvasive tool for SV estimation, addressing prior discrepancies and enhancing haemodynamic management in critical care.

Introduction

Accurate measurement of stroke volume is crucial for effective diagnosis and management of haemodynamics in cardiogenic shock. Traditionally, pulmonary artery catheterisation has been deemed the gold standard; however, its invasive nature and associated risks have sparked a demand for safer, noninvasive alternatives. In this context, echocardiography is vital for intensive care practitioners, potentially enhancing patient safety and care. Despite this, the reliability of various stroke volume estimation techniques remains inconsistent. Doppler-based estimations have shown a variable correlation with invasive measurements in critically ill patients.^1,2 Moreover, our previous studies have highlighted significant discrepancies in stroke volume estimations when comparing the Doppler technique to the modified Simpson’s method of discs. ³ These findings emphasise the urgent need to reassess the accuracy of noninvasive assessments to improve clinical outcomes in this vulnerable population.

Method

In this subgroup analysis of a recent study, ⁴ we examined the relationship between native stroke volume, measured by thermodilution via a pulmonary artery catheter, and simultaneously obtained left ventricular echocardiographic parameters in 39 patients with ischaemic cardiogenic shock without significant mitral regurgitation (moderate or more)

Details of the method can be found in the original study article. ⁴

Results

The patients had a median age of 61 years [52, 67], with 32 (83%) males. Mechanical circulatory support was required in 18 (46%) patients. Remarkably, stroke volume measurements obtained through left ventricular outflow tract (LVOT) Doppler and diameter (SV_Doppler) demonstrated a strong correlation by Spearman’s coefficient (r = 0.91, p < 0.0001) with the invasive stroke volume (Table 1). This robust relationship is supported by minimal systemic underestimation bias, quantified at only −3.6 ± 8 ml by the Bland-Altman test, and a clinically acceptable percentage error of 28% (Table 1, Figure 1). ⁵ Conversely, stroke volume derived using Simpson’s method (SV_Simpsons) displayed a weaker correlation (r = 0.71, p < 0.0001) with worse underestimation bias (−11 ± 14 ml) and a percentage error nearly doubling that of SV_Doppler (44%, Table 1). Furthermore, the ability of SV_Doppler to accurately predict invasive stroke volumes below 60 ml, with an area under the curve (AUC) of 0.91 in receiver operator curve analysis, underscores its superior diagnostic performance compared to SV_Simpsons (AUC = 0.84; Table 1 and Figure 1). These findings are especially pertinent given that the LVOT velocity time integral and the global longitudinal strain showed moderate correlations with invasive measurements, whereas ejection fraction did not correlate significantly (Table 1).

Table 1.

Echocardiography results and correlation of echocardiography parameters with the native stroke volume by thermodilution via pulmonary artery catheter in patients with ischaemic cardiogenic shock.

Parameter	Results	r-Value	p-Value	Percentage error	Mean difference (SD)	AUC	p-value	Sensitivity	Specificity
Stroke volume by PAC thermodilution, ml [IQR]	56 [40, 70]
LVEF, % [IQR]	31 [22, 35]	0.26	0.11
Stroke volume by Simpson’s method, ml [IQR]	43 [30, 59]	0.71	<0.0001	44%	-11 ml (±14)	0.84	0.0005	95%	47%
Stroke volume by Doppler echocardiography, ml [IQR]	56 [42, 66]	0.91	<0.0001	28%	-3.6 ml (±8)	0.91	<0.0001	100%	82%
LVOT VTI, cm [IQR]	17 [13, 20]	0.56	0.0003
LV GLS, % [IQR]	−8.3 [−10.3, − 6.2]	−0.56	0.001

AUC: area under the curve; IQR: inter-quartile range; LVEF: left ventricular ejection fraction; LV GLS: left ventricular global longitudinal strain; LVOT VTI: left ventricular outflow tract velocity time integral; PAC: pulmonary artery catheter; SD: standard deviation.

Figure 1.

Correlation between invasive stroke volume by thermodilution via pulmonary artery catheter and stroke volume estimation methods by echocardiography in patients with ischaemic cardiogenic shock: (a) correlation between invasive and Doppler stroke volume measurements, (b) correlation between invasive and Simpson’s method stroke volume measurements, (c) Bland-Altman graph for bias of Doppler stroke volume compared to invasive stroke volume, (d) Bland-Altman graph for bias of Simpson’s stroke volume compared to invasive stroke volume, (e) area under the curve (AUC) for Stroke volume by Doppler echocardiography prediction of reduced invasive stroke volume, and (f) AUC for Stroke volume by the modified Simpson’s method of discs prediction of reduced invasive stroke volume.

Discussion

These results consolidate the role of SV_Doppler as an accurate, noninvasive tool for clinicians managing patients with ischaemic cardiogenic shock in the absence of invasive monitoring.

It is essential to understand that obtaining accurate SV_Doppler readings depends on careful measurements of the LVOT diameter at the insertion points of the aortic valve leaflets, oriented perpendicularly to the long axis of the LVOT. ⁶ Moreover, the alignment of the Doppler sample volume with less than 20° deviation from the direction of blood flow in the LVOT is vital for precision. By strictly adhering to these technical standards as demonstrated in this study, we can significantly enhance the reliability of SV_Doppler readings, ultimately improving outcomes in real-time clinical applications.

Our study deliberately excluded patients with significant mitral regurgitation to uphold the integrity of our measurements and avoid discrepancies between the invasive stroke volume recorded in the right ventricle and the noninvasive measurements obtained through SV_Doppler in the left ventricle. Notably, differences between these measurements can provide critical insights into mitral regurgitation volume, significantly enhancing the diagnostic capability of this noninvasive tool and aiding in the identification of cardiogenic shock resulting from severe mitral regurgitation, where invasive stroke volume may appear adequate, yet left ventricular output remains compromised.

SV_Simpsons exhibited a weaker correlation with invasive measurements than SV_Doppler. This difference may arise from SV_Simpsons’ reliance on image quality, which declines with the severity of cardiogenic shock. ⁷ Furthermore, SV_Simpsons concentrates solely on specific regions of the left ventricle. Additionally, the underlying formula assumes that the left ventricle is ellipsoidal. ⁸ However, this assumption often does not always apply in the context of ischaemic cardiogenic shock, ⁹ underscoring a significant limitation of this method.

In conclusion, our findings reveal that SV_Doppler not only correlates strongly with invasive stroke volume measured by thermodilution through pulmonary artery catheterisation in patients experiencing ischaemic cardiogenic shock without significant mitral regurgitation but also suggests its potential as a vital therapeutic target for enhancing haemodynamic management in this critically ill patient population. This paves the way for further exploration of SV_Doppler’s capabilities in tailoring interventions in complex clinical scenarios.