Levosimendan in Septic Shock: the LeoPARDS study
“I want to be your medicine, I want to feed the sparrow in your heart …”
A 39 year old woman is admitted to the intensive care unit for hypotension, anuria and altered mentation despite 3 litres of intravenous lactated ringers infusion. She is febrile and found to have gram negative bacteremia of a urinary source. Her lactate remains 4.3 mmol/L with a mean arterial pressure of 63 mmHg whilst on norepinephrine and vasopressin infusions. Her urine output remains low and a bedside echocardiogram by the critical care fellow reveals a depressed ejection fraction. The possibility of levosimendan is raised on evening rounds.
Deep within the sulci of every intensivist’s cortex lies the accepted importance of inotropic support in the face of presumed septic cardiomyopathy. This almost certainly arises from the notion of central venous oxygen saturation as a part of the venerable early goal directed therapy algorithm for severe sepsis and septic shock. This, of course, was borrowed by Dr. Rivers from the goal directed therapy of the 1980s, using oxygen kinetics as a surrogate for cardiac output.
While dobutamine was classically used, excitement has swirled about the use of a novel inotrope, levosimendan. With this in mind, the authors of the LeoPARDS trial sought to feed the sparrow in the heart of those suffering from septic shock and - hopefully - reduce organ dysfunction as measured by the SOFA score.
Levosimendan is a calcium-sensitizing drug with a plethora of hemodynamic effects. In some respects, it is hemodynamically similar to dobutamine, however, levosimendan does not act through an adrenergic receptor. Levosimendan increases cardiac contractility and causes vasodilation; reportedly, these effects result in augmented myocardial contraction with only a small increase in myocardial oxygen demand. Further, levosimendan may have favourable lusitropic effects. Data of the benefit of levosimendan in patients – and in animal models – with organ dysfunction from sepsis have been mixed.
Adult patients from 34 general adult intensive care units [ICU]s were enrolled in the UK. The patients had to have both septic shock and received vasopressors for at least 4 hours in order to be included. Further, there could be no more than 24 hours between the time that the patient met the inclusion criteria and enrollment. Importantly, the APACHE II scores were slightly higher than recent studies of patients with life-threatening organ dysfunction as a consequence of infection [e.g. ProMISe, ProCESS, ARISE]. In the LeoPARDS trial, the average APACHE II score at enrollment was 25 in both arms.
Patients were randomly assigned to receive placebo or levosimendan; it was a double-blinded study. There was a dedicated algorithm for dosage changes and patients were otherwise managed based on the Surviving Sepsis Campaign guidelines. Fluids, vasopressors and inotropes were allowed at the discretion of the treating clinician.
The primary outcome was change in daily SOFA score and there was no statistically significant difference between placebo and levosimendan. However, there was a trend to lower SOFA score in placebo which may have been largely driven by a lower cardiovascular SOFA score in the placebo group; the difference was lessened when the cardiovascular system was excluded from the SOFA score. There was a non-statistically significant trend for increased 28 day mortality in those who received levosimendan as compared to placebo [34.5% versus 30.9% mortality at 28 days]. 8 patients in the levosimendan group suffered a supraventricular tachycardia as compared to 1 patient in placebo. Patients randomized to receive levosimendan were less likely than those in the placebo group to be successfully weaned from mechanical ventilation over the period of 28 days [HR of 0.77].
While certainly disappointing from a pharmacological perspective, the physiological implications of LeoPARD are tantalizingly nebulous. Why did this inotrope fail? As levosimendan bypasses the adrenergic receptor, there should have been less tachycardia and more oxygen delivery! In the supplementary data, there appears to be no difference in fluid administration between the two groups and cardiac index remains fairly similar. What’s interesting, however, is that the heart rate is notably higher in the early hours following randomization in the levosimendan group – despite a similar cardiac index. By implication, the stroke volume fell in the treatment arm. But why? This is impossible to know without continuous transesophageal monitoring, but we can hypothesize.
As stroke volume falls by either a fall in the end diastolic volume [EDV] or rise in the end-systolic volume [ESV] – or both – one would expect a fall in the EDV to be accompanied by diminished filling pressure. Using the central venous pressure [CVP] as a surrogate for left ventricular filling pressure [a stretch, I am aware], we see that there was no change between the two groups; those who received levosimendan or placebo had, on average, a CVP above 10 mmHg through the duration of the study.
This leaves us with a potential rise in end-systolic volume. As Dr. Burkhoff teaches us in his classic lecture on cardiac physiology [see figure 18], a rise in peripheral vascular resistance or heart rate will both increase the effective arterial elastance – the result being a rise in the ESV. Indeed, the higher percentage of patients on norepinephrine at higher doses with higher heart rates in the levosimendan group speak to this. In effect, any beneficial properties of heightened contractility from levosimendan were washed away by an increased need for norepinephrine with consequent higher heart rate and afterload. It appears, in totality, that the authors managed to concoct a fancy form of epinephrine with apparently more deleterious consequences – given that in the CATS trial dobutamine plus norepinephrine had a similar clinical outcome to epinephrine in the management of severe sepsis and septic shock.
One notable difference between epinephrine in the CATS trial and levosimendan in LeoPARDS is that the former reliably increased serum lactate despite increasing cardiac output. As lactate is thought to be an adaptive biochemical byproduct and augment cardiac contractility, the alluring question which effervesces and simmers just beneath the surface of the Surviving Sepsis sanity is the unthinkable – could lactate be a good thing? I know how Josh Farkas feels about this.
As to why levosimendan seemed to prolong liberation from mechanical ventilation, I'd love to read your thoughts in the comments below.