Understanding lactate in sepsis & Using it to our advantage

Once upon a time a 60-year-old man was transferred from the oncology ward to the ICU for treatment of neutropenic septic shock.  Over the course of the morning he started rigoring and dropped his blood pressure from 140/70 to 70/40 within a few hours, refractory to four liters of crystalloid.

In the ICU his blood pressure didn't improve with vasopressin and norepinephrine titrated to 40 mcg/min.  His MAP remained in the high 40s, he was mottled up to the knees, and he wasn't making any urine.  Echocardiography suggested a moderately reduced left ventricle ejection fraction, not terrible but perhaps inadequate for his current condition. 

Dobutamine has usually been our choice of inotrope in septic shock.  However, this patient was so unstable that we chose epinephrine instead.  On an epinephrine infusion titrated to 10 mcg/min his blood pressure improved immediately, his mottling disappeared, and he started having excellent urine output. 

However, his lactate level began to rise.  He was improving clinically, so we suspected that the lactate was due to the epinephrine infusion.  We continued the epinephrine, he continued to improve, and his lactate continued to rise.  His lactate level increased as high as 15 mM, at which point the epinephrine infusion was being titrated off anyway.  Once the epinephrine was stopped his lactate rapidly normalized.  He continued to improve briskly.  By the next morning he was off vasopressors and ready for transfer back to the ward.

This was eye-opening.  It seemed that the epinephrine infusion was the pivotal intervention which helped him stabilize.  However, while clinically improving him, the epinephrine infusion was also driving his lactate to very high levels.  How could this be?  Isn't lactate evil?  Isn't the entire point of sepsis resuscitation to normalize the lactate? 

The classical understanding of lactate in sepsis is flawed.  The following is a brief overview of newer ideas about lactate.  For a more complete discussion please see articles by Paul Marik listed below in the references. 

Traditionally it was believed that elevated lactate is due to anaerobic metabolism, as a consequence of inadequate perfusion with low oxygen delivery to the tissues.  This has largely been debunked.  Most patients with sepsis and elevated lactate have hyperdynamic circulation with very adequate delivery of oxygen to the tissues.  Studies have generally failed to find a relationship between lactate levels and systemic oxygen delivery or mixed venous oxygen saturation.  There is little evidence of frank tissue hypoxemia in sepsis.  Moreover, the lungs have been shown to produce lactate during sepsis, which couldn't possibly be due to hypoxemia (Marik 2014). 

This has significant implications for sepsis treatment.  Traditional belief in inadequate oxygen delivery led to multiple interventions to improve oxygen delivery (e.g. blood transfusion to target a hemoglobin of 10 mg/dL, use of inotropes to increased mixed venous oxygen saturation >70%, and nitroglycerine infusion for hypertensive patients).  Lack of oxygen deficiency may explain why these interventions have not proven to be beneficial.

Lactate elevation in sepsis seems to be due to endogenous epinephrine stimulating beta-2 receptors (figure below).  Particularly in skeletal muscle cells, this stimulation up-regulates glycolysis, generating more pyruvate than can be used by the cell's mitochondria via the TCA cycle.  Excess pyruvate is converted into lactate. 

This process is entirely aerobic, occurring despite adequate oxygen delivery.  Lactate generation doesn't occur because the mitochondria are unable to function in the absence of oxygen.  Instead, lactate generation occurs because the TCA cycle simply isn't able to keep up with a very rapid rate of glycolysis. 

Lactate serves as a metabolic fuel for the heart and brain in conditions of stress.  In a rat sepsis model, depletion of lactate caused cardiovascular collapse, which could be reversed by infusing sodium lactate (Levy 2007).  This study also found that selective blockade of beta-2 receptors decreased lactate levels and reduced survival duration.  In humans, RCTs have shown that concentrated sodium lactate improves cardiac output among post-CABG patients and heart failure patients (Nalos 2014, Leverve 2008)(1). 

Lactate correlates with illness severity, generally being a sign of badness.  This may lead to the misconception that lactate itself is harmful.  However, like sinus tachycardia, although elevated lactate is an ominous sign it still may function as a beneficial compensatory mechanism. 

This alternative understanding of lactate has some implications for bedside patient management. 

The autonomic nervous system and endogenous catecholamines are mysterious and confounding.  When exposed to the same infection, some patients have a weak endogenous catecholamine response and immediately develop hypotension.  Other patients have a robust release of endogenous catecholamines which supports their blood pressure, preventing hypotension (these are often younger patients who may look deceptively well). 

Lactate is a marker of endogenous catecholamine release (2).  This makes lactate useful for detecting patients who have occult shock:  patients who are maintaining their blood pressure due to a vigorous endogenous catecholamine response.  These patients may have deceptively reassuring vital signs, masking the fact that they are in a catecholamine-dependent shock state (simply using their own catecholamines rather than, for example, a norepinephrine infusion).  Elevated lactate identifies these patients as having an increased risk of death or decompensation, thus requiring more aggressive management.  Although most often associated with sepsis, occult shock with elevated lactate may be seen with any cause of shock. 

In 2010 Jones et al. demonstrated that trending serial lactate levels was non-inferior to using mixed venous oxygen saturation as a guide to sepsis resuscitation.  However, more recently the PROCESS, ARISE, and PROMISE trials have demonstrated that trending mixed venous oxygen saturation is unnecessary to begin with.  In hindsight, both interventions may be equally unnecessary. 

Currently it is unknown whether adding lactate to other resuscitation endpoints is beneficial.  For example, suppose a patient is doing well clinically (e.g. with an adequate blood pressure, good urine output, and down-titrating vasopressors) but has a persistently elevated lactate level.  Will escalating the resuscitation based on the lactate level be beneficial, or harmful due to over-resuscitation (e.g. volume overload, arrhythmic complications from vasopressors)? 

There is no clear evidence about how lactate might guide treatment intensity within the context of a modern sepsis resuscitation.  Many approaches are reasonable.  However, lactate is not an indicator of inadequate oxygen delivery, so an elevated lactate should not be blindly used as a trigger to increase oxygen delivery. 

The common fear of administering lactate reveals a misunderstanding of LR and the role of lactate in shock states.  First, LR contains sodium lactate (not lactic acid), and is therefore not acidotic.  Second, lactate probably has a beneficial role as discussed above (although it is very rapidly metabolized).  Occasional concern has been raised about the effect of LR on trending lactate levels, but this effect is minimal and the utility of precisely trending lactate levels is unclear. 

Unfortunately, Plasmalyte and Normosol were designed decades ago specifically to avoid the administration of lactate.  Their design was misguided, as discussed in further detail here.  For most critically ill patients, LR may be the best crystalloid. 

Epinephrine has been recommended as a second-line vasopressor by many authors including the Surviving Sepsis guideline.  Although popular abroad, it is rarely used in the US.  One common reason for avoiding epinephrine is concern that it may cause elevated lactate levels which could be harmful or confound serial trending of lactate.

Improved understanding of lactate may allow us to utilize epinephrine more often.  As discussed above, serial trending of lactate is of unknown value and should not dissuade us from using epinephrine if this is the best drug.  Elevated lactate levels might be beneficial and provide a dual action of epinephrine on the heart, rather than representing an undesirable "side-effect:"

In 2010 Wutrichexamined the prognostic value of changes in lactate following initiation of epinephrine infusion in patients with shock (mostly septic shock).  Survivors had significantly greater increases in lactate over the first four hours of epinephrine therapy compared to nonsurvivors (figure below). Thus, an epinephrine-induced rise in lactate may be a good prognostic sign, indicating that the epinephrine is working.

Epinephrine's properties may make it ideal for patients who fail to respond well to norepinephrine (+/- low-dose vasopressin).  Such patients often have adequate afterload, but need some additional inotropy.  At low doses (e.g. 0-10 micrograms/min), epinephrine functions as an inotrope (Moran 1993).  For patients who fail to respond to inotropic doses of epinephrine, higher doses of epinephrine will provide inotropy and vasoconstriction as well.  Thus, an epinephrine titration may be a simple approach to rapidly trial inotropic support and then provide additional vasoconstriction if needed. 

There is only one RCT comparing epinephrine vs. dobutamine as a second-line agent for patients with septic shock on norepinephrine (Mahmoud 2012).  These authors found that compared to dobutamine, epinephrine led to faster hemodynamic stabilization, greater urine output, higher lactate levels, and no mortality difference.  Unfortunately this study is very limited by the use of low doses of norepinephrine (0.1 mcg/Kg/min). 

One advantage of the norepinephrine-epinephrine combination is that it is difficult to screw up.  Epinephrine alone is generally adequate for septic shock (Myburgh 2008).  Therefore, any combination of norepinephrine and epinephrine is probably fine.  Alternatively, when patients end up on norepinephrine combined with dobutamine, it is easier to make significant titration errors (e.g. titrating off norepinephrine before dobutamine).