This is the new hotness in critical care discussions – the co-administration of intravenous esmolol to critically ill folks on high-dose vasopressor support in the ICU. It’s a fascinating thought – considering the alpha- and beta-stimulation of norepinephrine necessary to maintain central perfusion, co-administration of a sympatholytic seems self-defeating, in a sense.
However, this may not be the case. There are multiple dose-dependent effects of catecholamines on different tissues, as well as concern the tachycardia resulting from sympathomimetic myocardial stimulation in sepsis results in adverse outcomes. Based on prior observational evidence, these authors performed a randomized, open-label trial of esmolol co-administration in a cohort of critically ill patients on vasopressor support.
They randomized 154 patients with HR >95 and septic shock to esmolol, titrated to a HR between 80 and 94 bpm, vs. usual care. The hemodynamic variables showed, despite the use of esmolol, norepinephrine dosage was not significantly increased in order to maintain MAP. Improvements in stroke volume compensated for a lower heart rate, resulting in non-significantly lower cardiac index. However, what has everyone fascinated – the control group had 90% in-hospital mortality, compared with 67% for the esmolol group. I don’t think anyone disagrees this is statistically significant and clinically important.
There were differences in baseline variables between groups. The etiology of sepsis was substantially tilted towards peritonitis in the esmolol group, with obviously different causative organisms. Lactic acid, base excess levels, and SAPS II scores favored the esmolol group. The lead author and one co-author also have financial conflicts of interest with Baxter, the makers of esmolol (Brevibloc). Financial conflicts, open-label, and the size of the study all make me wary of reproducibility and the magnitude of the effect size.
So, there are problems warranting additional and independent confirmation. That said, the mortality benefit observed in this study is large enough I wouldn’t contest anyone who wanted to go ahead and start trying this in their highest-predicted mortality subset. However, I’d also consent patients/families to this therapy as experimental and prospectively collect data for at least retrospective pre-/post- comparisons.
“Effect of Heart Rate Control With Esmolol on Hemodynamic and Clinical Outcomes in Patients With Septic Shock A Randomized Clinical Trial”
www.ncbi.nlm.nih.gov/pubmed/24108526
Great post, thanks.
The first thing that strikes me is how unusual their resuscitation strategy was. Every patient had a swan and they titrated fluid to CVP and PCWP targets. Levosimendan was used rather than dobutamine. Such differences limit generalizability to practice in the US.
Goal-directed resuscitation is validated only for the initial hours of resuscitation (hence “early” goal-directed therapy). Continuing goal-directed resuscitation after these time points is of unclear value and may risk over-resuscitation. Morelli et al continued goal-directed resuscitation for 96 hours. It’s possible that the esmolol was helpful because it limited the iatrogenic harm of this prolonged resuscitative strategy.
The observation that esmolol improved cardiac function emphasizes a question of whether prolonged beta-agonism on the heart may cause calcium overload and myocardial dysfunction. It's unsafe to assume that the acute and chronic effects of beta-agonists will be the same.
Overall this is a seminal article, and I agree with their general argument that excess beta-agonism over days is detrimental. However, approaching this on a practical level, it may be easiest simply to de-escalate resuscitative goals after the early phase of resuscitation. It is possible that there are a subset of patients with prolonged vasoplegia following resolution of septic shock, who could be best served by neosynephrine.
Great post, thanks.
The first thing that strikes me is how unusual their resuscitation strategy was. Every patient had a swan and they titrated fluid to CVP and PCWP targets. Levosimendan was used rather than dobutamine. Such differences limit generalizability to practice in the US.
Goal-directed resuscitation is validated only for the initial hours of resuscitation (hence “early” goal-directed therapy). Continuing goal-directed resuscitation after these time points is of unclear value and may risk over-resuscitation. Morelli et al continued goal-directed resuscitation for 96 hours. It’s possible that the esmolol was helpful because it limited the iatrogenic harm of this prolonged resuscitative strategy.
The observation that esmolol improved cardiac function emphasizes a question of whether prolonged beta-agonism on the heart may cause calcium overload and myocardial dysfunction. It's unsafe to assume that the acute and chronic effects of beta-agonists will be the same.
Overall this is a seminal article, and I agree with their general argument that excess beta-agonism over days is detrimental. However, approaching this on a practical level, it may be easiest simply to de-escalate resuscitative goals after the early phase of resuscitation. It is possible that there are a subset of patients with prolonged vasoplegia following resolution of septic shock, who could be best served by neosynephrine.
Great appraisal – I hadn't spotted the link with Baxter
Great appraisal – I hadn't spotted the link with Baxter
Great review as usual. Thanks for pointing out the link w/ pharma. However, I was surprised to read your comment that "I don't think anyone disagrees this is statistically significant and clinically important" regarding the mortality benefit. You typically are very strict with your interpretations especially since this wasn't a primary outcome and the study wasn't powered to detect this. I agree it seems intriguing and your later comments do somewhat reflect that. Though not sure this is ready for anything except within study environment.
Great review as usual. Thanks for pointing out the link w/ pharma. However, I was surprised to read your comment that "I don't think anyone disagrees this is statistically significant and clinically important" regarding the mortality benefit. You typically are very strict with your interpretations especially since this wasn't a primary outcome and the study wasn't powered to detect this. I agree it seems intriguing and your later comments do somewhat reflect that. Though not sure this is ready for anything except within study environment.
You're right – this is a secondary endpoint, so strictly speaking it requires specific interventional confirmation. The trouble with knowledge translation, however, is there are not always black & white interpretations of the evidence. There are many flaws with this study – but they show a massive mortality benefit. Mortality, as an endpoint – even secondary – is about as clinically important as they get.
Therefore, in the highest mortality groups that seem to best match the population studied, it's not unreasonable to base treatment decisions on the observed statistically significant and clinically important secondary mortality outcome.
You're right – this is a secondary endpoint, so strictly speaking it requires specific interventional confirmation. The trouble with knowledge translation, however, is there are not always black & white interpretations of the evidence. There are many flaws with this study – but they show a massive mortality benefit. Mortality, as an endpoint – even secondary – is about as clinically important as they get.
Therefore, in the highest mortality groups that seem to best match the population studied, it's not unreasonable to base treatment decisions on the observed statistically significant and clinically important secondary mortality outcome.