Evolution of haemodynamics and outcome of fluid-refractory septic shock in children

Abstract

Background: Maintaining threshold values of cardiac output (CO) and systemic vascular resistance (SVR) when used as part of the American College of Critical Care Medicine (ACCM) haemodynamic protocol improves the outcomes in paediatric septic shock.

Objective: We observed the evolution of CO and SVR during the intensive care admission of children with fluid-refractory septic shock and report this together with the eventual outcomes.

Design: Prospective observational study.

Setting: Tertiary care Paediatric Intensive Care Unit (PICU) in London.

Methods: Children admitted in fluid refractory septic shock to the Intensive Care Unit over a period of 36 months were studied. Post liver re-transplant children and delayed septic shock admissions were excluded. A non-invasive ultrasound cardiac output monitor device (USCOM) was used to measure serial haemodynamics. Children were allocated at presentation into one of two categories: (1) hospital-acquired infection and (2) community-acquired infection. Vasopressor, inotrope or inodilator therapies were titrated to maintain threshold cardiovascular parameters as per the ACCM guidelines.

Results: Thirty-six children [19 male, mean age (SD) 6.78 (5.86) years] were admitted with fluid-refractory septic shock and studied. At presentation, all 18 children with hospital-acquired (HA) sepsis and 3 from among the community-acquired (CA) sepsis group were in 'warm shock' (SVRI < 800 dyne s/cm(5)/m(2)) whereas 15 of the 18 children with community-acquired sepsis and none in the hospital-acquired group were in 'cold shock' [cardiac index (CI) < 3.3 l/min/m(2)]. All 21 children in 'warm shock' were initially commenced on a vasopressor (noradrenaline). Despite an initial good response, four patients developed low CI and needed adrenaline. Similarly, all 15 children in cold shock were initially commenced on adrenaline. However, two of them subsequently required noradrenaline. Five others needed milrinone as an inodilator. In general, both groups of children had normalised SVRI and CI within 42 h of therapy but required variable doses of vasopressors, inotropes or inodilators in a heterogeneous manner. The overall 28-day survival rate was 88.9 % in both groups. Central venous oxygen saturation (ScvO2) was significantly (p = 0.003) lower in the community-acquired group (mean 51.72 % ± 4.26) when compared to the hospital-acquired group (mean 58.72 % ± 1.36) at presentation but showed steady improvement during therapy. Gram-positive organisms were predominant in blood cultures, 61 % in HA and 56 % in CA groups.

Conclusions: In general, we found children with community-acquired septic shock presented in cold shock whereas hospital-acquired septic shock children manifested warm shock. Both types evolved in a heterogeneous manner needing frequent revision of cardiovascular support therapy. However the 28-day survival in both groups was the same at 89 %. Frequent measurements of haemodynamics using non-invasive ultrasound helped in fine tuning cardiovascular therapies.