July 26, 2019

Optimise your therapeutic strategy

Cardiovascular clusters in septic shock combining clinical and echocardiographic parameters: a post hoc analysis


Better characterisation of circulatory failure in septic shock patient could help to optimise haemodynamic support. In this paper Geri et al reconsider the classic well-known path of haemodynamic alterations in septic shock with:

1 early phase with a low flow state related to hypo-volemia,

2 second phase after the initial resuscitation with a hyperdynamic state,

3 third phase with cardiac failure leading to multi-organ failure and death


Indeed, myocardial alteration is now known to possibly be present early in septic shock, with some studies emphasising a pretty high incidence (39%) of LV systolic dysfunction during the first day in ICU.

Therefore, authors have merged two published prospective databases for haemosepsis and haemopred study, with 12 different ICUs, including echocardiographic monitoring performed by a trans-esophageal route at the initial phase of septic shock for a post hoc analysis.

360 patients were analysed.  11.7% of patients were already receiving inotropic drugs at inclusion (dobutamine or epinephrine,). Day 7 and in-ICU mortality were 20.1% and 35%, respectively.

Five distinct cardiovascular phenotypes were identified through cluster analysis in this population.

> “Well resuscitated” (Cluster 1 = 16.9%) = neither LV systolic dysfunction and RV failure nor fluid responsiveness

> « LV systolic dysfunction” (Cluster 2 = 17.7% of patients) exhibiting low LVEF, LVFAC, and CI with high lactate and not fluid respondes.

> “hyper-kinetic” state (Cluster 3 = 23.3%) LV systolic function and CI were increased with no sign of fluid responsiveness.

> “RV failure” (Cluster 4= 22.5%) high RV/LV EDA ratio with a normal LV systolic function no more fluid responsive.

> “Still hypovolemic” (Cluster 5 = 19.4%) low CI, despite an increased LV systolic function (LVEF) due to sustained fluid responsiveness, as reflected by markedly elevated ∆SVC and low preload (CVP).


ICU mortality was 21.3 [95% CI 13.0; 33.1], 50.0 [38.1; 61.9], 23.8 [16.0; 33.9], 42 [31.8; 52.8], and 38.6 [28.0; 50.3] % in Clusters 1, 2, 3, 4, and 5.

A simple clustering approach combining echocardiographic parameters (LVEF, LVFAC, aortic VTI, RV/LV EDA, ∆SVC, mitral E wave velocity, and E′ wave velocity) and clinical parameters (heart rate, blood pressure, type and dose of catecholamine) allowed authors to characterise five distinct cardiovascular phenotypes.



Clinical features of those patient characterised in 5 cardiovascular phenotypes is well discussed. Is high afterload secondary to noreprinephrine infusion involved in septic cardiomyopathy with LV dysfunction?

Could this better characterisation of circulatory failure in septic shock help physicians to individualise the haemodynamic support for those patients who need inotrops, optimise haemodynamic and respiratory support for patients with RV failure, to go on with fluids etc.?

This study however was focused only on a haemodynamic evaluation by CCE, obtained in the first 12 hours after initial resuscitation of septic shock patients.

It might be of interest to assess in future studies whether the transition from one cardiovascular phenotype to another during the first 2–3 days could alter outcome, and whether it is related to spontaneous progression or induced by therapy.

Patient inclusion was not based on the Sepsis 3 definition of Septic shock.



Using a clustering approach in a large cohort of patients with septic shock evaluated early by CCE, five distinct cardiovascular phenotypes were revealed:

“Well resuscitated”, « LV systolic dysfunction”, “hyperkinetic” state, “RV failure” and “Still hypovolemic”.

Have an early and close look at echocardiographic parameters (LVEF, LVFAC, aortic VTI, RV/LV EDA, ∆SVC, mitral E wave velocity, and E′ wave velocity) and clinical parameters (heart rate, blood pressure, type and dose of catecholamine) to identify in which group your patients are and possibly optimise your therapeutic strategy.


This article review was prepared and submitted by Paul Abraham on behalf of the ESICM Journal Review Club.



1) Vieillard-Baron A, Caille V, Charron C et al (2008), Actual incidence of global left ventricular hypokinesia in adult septic shock. Crit Care Med 36:1701–1706.

2) Vieillard-Baron A, Prin S, Chergui K et al (2003), Hemodynamic instabil‑ ity in sepsis: bedside assessment by Doppler echocardiography. Am J Respir Crit Care Med 168:1270–1276.

Comment on this news