Are usual vasopressor doses causing unrecognized harm?
Two meta-analyses suggest "permissive hypotension" might save lives in vasodilatory shock
Maintaining a mean arterial pressure (MAP) of 65 mm Hg has been advised for patients with sepsis and other forms of vasodilatory shock. In practice, treating teams often exceed this target by a significant margin, such as in patients with more severe organ failure or pre-existing hypertension. This strategy is likely pursued out of concern that lower MAP targets would produce suboptimal perfusion in some (or most) patients.
More than a dozen randomized controlled trials (RCTs) have sought to identify whether higher or lower MAP targets produce better outcomes in critically ill patients with shock.
The following table summarizes three high-quality RCTs performed in the past ten years, enrolling 3,352 patients at 103 hospitals in the U.K., France, and Canada:
We reviewed the SEPSISPAM study back in 2014:
And the 65 trial in 2020:
These three trials shared their patient-level data (as opposed to the usual practice of only sharing summary results in published trials).
This allowed for a higher-resolution meta-analysis to be subsequently performed by Angriman et al (NEJM Evidence 2024).
In that meta-analysis, patients assigned to a lower MAP target had a risk ratio of 0.93 for 90-day mortality that did not reach statistical significance (95% confidence interval 0.76 to 1.07, with low certainty). This 7% relative risk reduction would represent about a 2% absolute reduction in mortality.
Their Bayesian analysis estimated an 87% posterior probability of benefit from lower MAP targets; authors state the effect persisted when using pessimistic priors and was robust to other sensitivity analyses.
No patient subgroup clearly benefited from a higher MAP target (or was harmed by a lower MAP), including patients with sepsis, pre-existing hypertension, or higher organ failure scores. However, confidence intervals were wide and could not rule out benefit or harm.
In another meta-analysis, D’Amico et al (Crit Care Med 2024) combined 12 RCTs enrolling critically ill patients, and 16 in perioperative settings (total N=15,672), concluding that patients in the lower MAP groups had:
About 1% absolute lower mortality risk (7% relative risk reduction; RR 0.93, statistically significant)
Lower rates of atrial fibrillation
Fewer transfusion requirements
The positive findings were driven mainly by patients in the ICU setting.
The two meta-analyses assessed very different patient populations, as demonstrated by their observed mortality rates (30-45% in the studies included in the Angriman meta-analysis where sepsis predominated, vs. 13-14% in the trials forming the D’Amico meta-analysis).
Both meta-analyses found a 7% relative risk reduction in mortality among those treated with a lower MAP target.
Conclusions
Targeting a lower MAP target could be beneficial in patients with vasodilatory shock, which is most often due to sepsis.
Any potential benefit would appear to be small. Because septic shock is so common, though, on a population level such an effect could theoretically scale to imply thousands of lives saved or lost each year depending on the MAP target chosen.
A randomized trial sufficiently powered to detect a 7% relative reduction in mortality (e.g., an absolute 35% reduced to 32.5%) at 80% power and alpha of 0.05 would require about 11,000 patients (per ClinCalc); no such trial is currently in process or planned on clinicaltrials.gov.
Owing to physiologic lability in critical illness, adopting a practice of strictly targeting a mean arterial pressure of 60-65 mm Hg would create a situation where many patients’ MAPs would frequently fall below the threshold currently considered safe. Although it’s possible this would be beneficial, prevailing care standards and clinician anxiety will resist the diffusion of “permissive hypotension” in vasodilatory shock.
It’s unlikely that these meta-analyses will meaningfully change that state of affairs, but they may nudge intensivists toward lower MAP targets in patients with sepsis or other causes of vasodilatory shock.
I always felt a mean of 65 may be way too much, especially since the diastolic is so low in distributive shock. I think targeting a systolic of 90 is enough. This is great, and I will plan on podcasting about this.
This is such a wonderful post. When I saw the email that you had a new post, the dopamine hit was immediate. Given the current state of the world, your posts are the rock of my salvation. Thank you for all you do. You have your fans, and your fans are grateful!