Written by Samuel Rouleau
Spoon Feed
In patients with sepsis, new LV dysfunction is associated with in-hospital and ICU mortality.
Broken hearts portend badness in sepsis
In this multi-center prospective study, adult patients with sepsis (n=2,274), who underwent screening echocardiography and did not have pre-existing congestive heart failure, were placed into 3 groups based on ejection fraction: normal (LVEF > 50%, n=1,803), moderate dysfunction (LVEF 30-50%, n=356), and severe dysfunction (LVEF < 30%, n=115). Multiple propensity matching models were used to address potential confounders.
In-hospital mortality was statistically different across the groups, with mortality rate of 40% in those with severe LV dysfunction, 29.5% in those with moderate, and 25.7% in those with none. Volume of fluid resuscitation was similar across groups, although rates of vasopressor requirement, mechanical ventilation, and renal replacement therapy were highest in those with severe LV dysfunction. The incidence of cardiac arrest was also highest in the severe LV dysfunction group (OR 4.15, 95%CI 2.29-7.52). Within those with severe LF dysfunction, history of prior cardiovascular disease was actually a protective factor, possibly related to myocardial ischemic conditioning.
How does this change my practice?
Concurrent septic shock and LV dysfunction confers high morbidity. This study offers some key insights:
- Fluid resuscitate as indicated, even when LV function is depressed.
- Bacteremia is associated with higher mortality rates with concurrent LV dysfunction; early source control is key.
- In general, I am hypervigilant in critically-ill septic patients with LV dysfunction. As their capillary leak syndrome resolves, they can easily tip into cardiogenic shock.
Source
Association Between Initial Left Ventricular Systolic Dysfunction and Clinical Outcome in Sepsis: A Multicenter Cohort Study. Crit Care Med. 2025 Jul 2. doi: 10.1097/CCM.0000000000006771. Epub ahead of print. PMID: 40601382
I’d love to hear more about the idea that we should fluid resuscitate as indicated even when the LV function is depressed. It seems to me like the three groups got similar amounts of fluid but the severely depressed group did much worse overall. My worry is that if we give fluids to patients with severe LV dysfunction then we could potentially make them worse and increase the chance that they needs things like pressors, mechanical ventilation and RRT. It may be that LV dysfunction is the only real risk factor for doing worse but I worry that a study like this can’t give us a clear answer in terms of the actual risk posed by a relatively liberal use of fluids.
Hi Matthew,
Thanks for the great comment! I appreciate your thoughtful approach to fluid resuscitation. The bottom line is that giving too much fluid to patient’s with LV dysfunction is harmful. I think the field has recognized the risks of volume overloading patients with LV dysfunction in sepsis, and initiation of early vasoactive therapy is common and beneficial. When I wrote “fluid resuscitate as indicated,” my intention is for the clinician to fluid resuscitate as indicated — do not blindly give 30 mL/kg!
Before going into some more detail, it is important to recognize that it is probably impossible to come up with a standardized fluid resuscitation guideline/goal for this sub-group of patients. Here is a nice editorial from 2020 in Crit Care that talks about the intricacies and conflicting evidence about fluid resuscitation in septic patients with heart failure: https://pmc.ncbi.nlm.nih.gov/articles/PMC7984412/.
Whenever I have a patient in shock, I go through the following schema to make sure I am addressing all aspects of their resuscitation.
BP = CO x SVR
SVR – always low in septic shock, support with vasoactives (in the ED, we typically start with norepi +/- vasopressin)
CO (factors listed below) = SV x HR
Preload – while some argue that vasoactives increase the “stressed” volume, when there is ongoing capillary leak like in septic shock, fluid resuscitation is still necessary to achieve an effective intravascular circulatory volume. Cranking up SVR in a patient without proper filling pressures will not fix shock.
Afterload – SVR and afterload are often lumped together. SVR is a component of afterload. Afterload also includes arterial compliance, aortic valve factors (think stenosis), or other etiologies that result in obstructing to flow. In septic shock, afterload is typically low. Yet, I think it about it separately, because I will also do the mental exercise of thinking about valvular pathology here or LVOTO, which is probably under recognized.
Inotropy – is the patient in mixed shock? Do they need an inotrope for maybe some low dose epi to augment. In early septic shock, many patients with LV dysfunction might not need inotropic support as the afterload reduction from their pathologic decrease in SVR may assist with forward flow. However, I find that when these patients with reduced LV function are in the stabilization phase and their septic shock is improving, they are high risk for tipping into cardiogenic shock when they start mobilizing fluid and/or need inotropic support to assist with diuresis.
Rate – can be too fast or too slow. For instance, a patient who is pacemaker dependent may need their paced rate turned up. I recently had a patient in septic shock from pneumonia and an empyema. Turned up their pacemaker from a set rate of 60 to 80, and their MAP increased by 10! Alternatively, if the HR is too high, then stroke volume might be low due to inappropriate diastolic filling time (especially in AFib with RVR, HOCM & LVOTO as other examples). In some of these situations, it is worthwhile using rate control to better control HR to increase diastolic filling time and therefore stroke volume.
Rhythm – Discussed a bit above in the rate discussion. If there are arrythmias that are of hemomodynamic significance, be aggressive about fixing them!
Valvular pathology – all of the above becomes way more complicated in those with serious valve issues, but make sure you’re not missing a new diagnosis of a critical valvulopathy.
Endocrinopathies when severe will also impact cardiac output and should be considered.
Ok – back to the main point now :). Anecdotally, I have been consulted for admission on numerous patients with profound septic shock and a concurrent medical condition that typically precludes aggressive volume resuscitation (ESRD, liver cirrhosis, LV dysfunction, RV dysfunction), and the ED clinicians have been appropriately aware of these indications to limit volume, but are not always doing thorough hemodynamic assessment. In some of these instances, the patients will be on really high doses of vasoactives with only a 250 mL fluid bolus and getting sicker. One of my first moves (after doing an evaluation outlined above) is add on some more fluid resuscitation while being judicious and mindful of how I am doing it and what my specific resuscitation goal is. While volume overload is harmful, under fluid resuscitation in septic shock is also harmful, and I sometimes wonder if the pendulum has swung a bit too far.
So – yes! Fluid resuscitate as indicated in patients with septic shock and LV dysfunction. These are some of the sickest patients in the hospital, and they deserve a thorough assessment of their hemodynamics with targeted resuscitation to get an optimal outcome.
Thanks for the really insightful response. I do worry that at times we are so worried about giving folks too much fluid that we overlook the patients who actually need more volume. Like lots of things in medicine I think the pendulum has swung maybe a little bit too far in some instances. Thanks again for breaking it down so nicely!