This is the first of a series of four posts about vasopressors and hemodynamics. The first step is to explore the physiology of a pure alpha-1 agonist (phenylephrine).
Make no mistake, I'm not very fond of phenylephrine. I rarely use it (mostly for hypotensive atrial fibrillation). However, understanding phenylephrine is a prerequisite to understanding related vasopressors, particularly midodrine and norepinephrine.
Evidence regarding phenylephrine consists of a patchwork of often contradictory human and animal studies, based mostly on surrogate endpoints. Thus, this post is framed as an alternative viewpoint: a perspective which remains debatable.
Traditional belief #1: Norepinephrine is awesome, but phenylephrine is evil.
Alternative viewpoint: Norepinephrine and phenylephrine have more similarities than differences
There is a misconception that norepinephrine is dramatically different from phenylephrine. However, the pharmacology of these drugs is similar. Norepinephrine is equivalent to phenylephrine plus a bit of beta-stimulation.
Evidence: Studies comparing phenylephrine vs. norepinephrine
Morelli 2008a randomized 32 patients with hyperdynamic sepsis to receive phenylephrine vs. norepinephrine as a first-line vasopressor (with additional open-label dobutamine as needed). The only difference detected between the groups was that patients in the norepinephrine group had a slightly higher blood pressure. There were no differences in heart rate, cardiac output, systemic vascular resistance, lactate, gastric mucosal perfusion, renal function, or dobutamine requirement:
Morelli 2008b and Reinelt 1999 performed crossover studies in patients with septic shock who were transitioned from norepinephrine to phenylephrine, and then back to norepinephrine. Both studies found nearly identical hemodynamics with either drug (the only difference was a slightly lower heart rate in the phenylephrine group in Morelli 2008b, table shown below)(1).
Kee 2015 performed a RCT comparing norepinephrine vs. phenylephrine during spinal anesthesia for cesarean delivery of 104 women. Phenylephrine maintained a stable cardiac output whereas norepinephrine caused a slight increase in the cardiac output (figures below). Patients on either drug had similar stroke volumes, with differences in cardiac output driven by differences in heart rate.
Poterman 2015 performed a RCT comparing norepinephrine vs. phenylephrine among sixty patients undergoing anesthesia with propofol and remifentanil. The physiologic effects of initiating either drug were indistinguishable (2):
Making sense of this data?
Patients may have variable responsiveness to the relatively weak beta-adrenergic effects of norepinephrine. For example, responsiveness to beta-adrenergic stimulation may be lower in patients with sepsis (in whom cardiac beta-receptors may be down-regulated) or patients with a robust endogenous beta-adrenergic activity (in whom exogenous norepinephrine may have little additional effect).
Thus in some contexts, the beta-agonist activity of norepinephrine seems to have no clinical effect (e.g. Morelli 2008a, Reinelt 1999, and Poterman 2015 above). In these situations, norepinephrine and phenylephrine act in an identical fashion. Alternatively, in Kee 2015, norepinephrine did have measurable beta-agonist effects, leading to an increase in heart rate and cardiac output compared to phenylephrine.
Traditional belief #2: Phenylephrine reduces the cardiac output.
Alternative viewpoint: Phenylephrine has a variable effect on cardiac output
Most textbooks and review articles contain a table showing a reduction in cardiac output from phenylephrine (example above). The usual explanation is that phenylephrine increases systemic vascular resistance (afterload), which makes it harder for the heart to pump blood forward. Reality is more complicated.
Phenylephrine can affect cardiac output in several ways:
- Reflex bradycardia: Increased blood pressure may trigger the carotid baroreflex, causing a decreased heart rate which reduces cardiac output.
- Venoconstriction: Constriction of veins increases venous return (i.e. the preload), potentially increasing cardiac output.
- Arterial constriction: Constriction of systemic arteries increases the systemic vascular resistance (i.e. the afterload), potentially reducing cardiac output.
- Baseline heart rate and autonomic tone: For a patient who is severely tachycardic (e.g. atrial fibrillation), reducing the heart rate could be beneficial. Alternatively, in patients with very high endogenous adrenergic tone driving a sinus tachycardia, phenylephrine could have less effect on heart rate.
- Preload-responsiveness: Venoconstriction will serve to increase the cardiac output only in patients who are preload-responsive (i.e. fluid-responsive).
- Afterload responsiveness: Patients with systolic heart failure may respond to increased afterload with a decrease in stroke volume, whereas patients with greater cardiac reserve function may maintain their stroke volume.
Depending on the balance of these factors, phenylephrine can have any effect on cardiac output (increased, decreased, or unchanged). This was demonstrated by Yamazaki 1982, who administered phenylephrine to either patients with hyperdynamic septic shock or heart failure. Patients with heart failure experienced a decrease in cardiac output with phenylephrine (adjacent figure). These patients may not have been preload-responsive due to their cardiac disease. Thus, the increase in afterload had a dominant effect, causing their cardiac output to decrease.
Patients with hyperdynamic sepsis responded differently. These patients didn’t experience an increase in afterload at the dose of phenylephrine used, possibly due to systemic vasoplegia from sepsis. The dominant effect in these patients was an increase in preload, which caused an increase in cardiac output.
Additional studies confirm that phenylephrine usually doesn't reduce cardiac output in hyperdynamic sepsis. Flancbaum 1997 found cardiac output to be stable in septic patients despite titrating phenylephrine up to 8 ug/kg/min (table above). Gregory 1991 found that phenylephrine produced a stable/increased cardiac output, with increased stroke volume:
Traditional belief #3: Phenylephrine reduces the cardiac output by increasing afterload, leading to a reduced stroke volume.
Alternative viewpoint: Reduced heart rate might be the most common mechanism of decreased cardiac output
There are clearly situations where phenylephrine reduces the cardiac output. This has traditionally been attributed to increased afterload causing a reduction in stroke volume. Reduced stroke volume has been shown after phenylephrine boluses and in patients with heart failure (who may be unable to tolerate additional afterload, Yamazaki 1982)(3).
However, in most studies reduced cardiac output from phenylephrine infusions occurs due to reflex bradycardia (without any reduction in stroke volume; Bellomo 2012). For example, Soeding 2013 described the effect of phenylephrine versus fluid loading in patients undergoing upright positioning for shoulder surgery. As shown here, phenylephrine caused a parallel reduction in heart rate and cardiac index, without reducing fractional shortening (a measurement of ejection fraction).
This is usually a subtle effect that easily goes unnoticed. For example, a patient's heart rate might drop from 87 b/m to 70 b/m, causing their cardiac output to fall by 20%. A heart rate of 70 b/m isn't technically “abnormal,” so it won't draw attention. Nonetheless, this heart rate may be suboptimal for a shocked patient (“abnormally normal”).
Effects on heart rate may be a useful clinical sign, because they are easily observable without sophisticated hemodynamic monitors. An inappropriately low heart rate in a patient on phenylephrine might suggest that cardiac output is being suppressed by reflex bradycardia (4).
Studies investigating cardiac output changes associated with phenylephrine suggest that heart rate changes parallel changes in cardiac output. –Habib 2012
Pre-existing tachycardia changes this. Jain 2010 randomized septic patients refractory to high-dose dopamine to receive norepinephrine or phenylephrine. At baseline, patients had an average heart rate of 150 b/m from the dopamine. Compared to norepinephrine, phenylephrine caused a decreased heart rate, an increase in stroke volume (better diastolic filling), and no reduction in cardiac output. This supports using phenylephrine for patients with hypotension and severe tachycardia (e.g. hypotensive atrial fibrillation).
Traditional belief #4: Phenylephrine reduces renal blood flow and hurts the kidneys.
Alternative viewpoint: Phenylephrine may be beneficial to the kidneys in some situations.
Phenylephrine is widely believed to reduce renal blood flow and thereby impair renal function. However, phenylephrine doesn't vasoconstrict the renal vasculature more than other vascular beds. Thus, it is possible that the effect of phenylephrine on renal blood flow, similar to the effect on cardiac output, may be context-dependent. For example, Bellomo 2012 found that phenylephrine increased renal blood flow both in normal and septic sheep.
More importantly, it is incorrect to assume that renal function changes in parallel with renal blood flow. For example, septic patients may have shunting of blood through the kidneys, with renal failure despite increased total renal blood flow. Drugs that vasoconstrict the efferent arterioles (e.g. vasopressin) may improve renal function despite causing a reduction in renal blood flow. Therefore, instead of focusing on renal blood flow it is better to focus on changes in renal function (e.g. glomerular filtration rate, urine output).
Available evidence suggests that phenylephrine improves renal function. In healthy female volunteers, phenylephrine increased urine output and glomerular filtration rate (Klein 1995). Gregory 1991 found that phenylephrine increased urine output in patients with septic shock.
Other alpha-1 agonists (midodrine and methoxamine) seem to improve renal function. Midodrine is widely used for treatment of renal failure in the context of hepatorenal syndrome. Sun 2014 found that methoxamine increased intraoperative urine output in a prospective RCT involving patients undergoing elective hip replacement.
Norepinephrine has a good track record for defending renal perfusion. How does phenylephrine compare? Morelli 2008a and Jain 2010 found no difference in urine output in RCTs comparing phenylephrine vs. norepinephrine in septic shock.
- Norepinephrine is similar to phenylephrine (norepinephrine is equivalent to phenylephrine plus some beta-1 stimulation). In some situations the two drugs may function in an identical fashion.
- Like norepinephrine, phenylephrine can cause venoconstriction, thereby increasing venous return to the heart.
- Phenylephrine can affect cardiac output in a variety of ways:
- Increased preload may increase cardiac output
- Increased afterload and reflex bradycardia may reduce cardiac output.
- The effect of phenylephrine on cardiac output varies depending on the clinical context.
- The most common mechanism whereby phenylephrine reduces cardiac output may be reflex bradycardia. Thus, an inappropriately low heart rate in a shocked patient on phenylephrine suggests that cardiac output is being suppressed.
- Available evidence suggests that phenylephrine can improve renal function, particularly in hyperdynamic shock states (e.g. cirrhosis, sepsis).
Stay tuned for three additional posts on hemodynamics.
- Vasopressor basics (EMCrit)
- The crashing atrial fibrillation patient (EMCrit)
- Decompensation of atrial flutter following cardioversion (PulmCrit)
- Collapsed IVC doesn't equal hypovolemia (PulmCrit)
- Early initiation of norepinephrine in septic shock (PulmCrit)
- Vasopressin and renal microvascular hemodynamics (PulmCrit)
- Morelli 2008b found a reduced GFR in the phenylephrine group, but this remained low after transitioning back to norepinephrine. Thus, it is not possible to conclude that the low GFR was caused by phenylephrine, it may have simply represented the evolution of septic shock over time. Phenylephrine did reduce the hepatic perfusion compared to norepinephrine, but it is unclear exactly what the hepatic perfusion ideally should be, and whether reduced hepatic perfusion could be a beneficial or detrimental effect.
- The cardiac output data from this study was based on pulse contour analysis using a noninvasive monitoring device (Nexfin), whose validity is unclear. However, reported heart rates should be accurate.
- The physiologic effect of a phenylephrine bolus is not necessarily the same as a phenylephrine infusion. For example, Cannesson 2012 found in pigs that when volume loaded (and preload-independent), a phenylephrine bolus reduced cardiac output. However, in this same setting a phenylephrine infusion had a positive/neutral effect on cardiac output. The response to a phenylephrine bolus represents the body's initial response, but may not represent the new equilibrium state that is eventually reached when ongoing phenylephrine exposure occurs in the form of an infusion. This post is focused on phenylephrine infusions. In general it seems that the response to phenylephrine boluses is less favorable than infusions. It is possible that studies of phenylephrine boluses may have contributed to phenylephrine's bad reputation.
- On a more practical and less arcane level, a shocked patient with a sluggish heart rate who is perfusing poorly could probably benefit from a faster heart rate, regardless of the exact etiology of the heart rate (i.e. phenylephrine, beta-blocker, sinus node disease, etc.). We spend lots of time thinking about inotropy, but chronotropy may also be important. For example, if you double someone's heart rate from 40 b/m to 80 b/m with a stable stroke volume, their cardiac output will double. Meanwhile, it is impossible to double someone's cardiac output by increasing their ejection fraction from 55% to 110%.
- IBCC chapter:Guide to APRV for COVID-19 - April 8, 2020
- PulmCrit Theoretical Post – The COVID Severity Index (CSI 1.0) - April 2, 2020
- PulmCrit wee – Why the SCCM/AARC/ASA/APSF/AACN/CHEST joint statement on split ventilators is wrong. - March 29, 2020