Sunday, March 23, 2014

Lost Without a MAP: Blood Pressure Targets in Septic Shock

Another of the critical care articles published early online at last week was this trial of High versus Low Blood-Pressure Target in Patients with Septic Shock.  In this multicenter, open-label trial, the authors enrolled 776 patients in France and randomized them to a target MAP (mean arterial pressure) of 65-70 mm Hg (low target) versus 80-85 (high target).  The hypothesis is that a higher pressure, achieved through vasopressor administration, will improve 28-day mortality.  If you don't already know the result, guess if the data from this trial support or confirm the hypothesis (the trial had 80% power to show a 10% absolute reduction in mortality).

Well, if this isn't your first visit to the Medical Evidence Blog, you know that it is highly unlikely that anything we do in critical care or elsewhere is going to have an effect on 28-day mortality, especially when the study is powered for a 10% absolute mortality reduction.  And that was the case here:  mortality at 28 days was 36.5% in the high target group and 34.0% in the low target group.  There were likewise no strong trends in favor of either group in 90 day mortality, or the secondary outcomes of need for mechanical ventilation, LOS (length of stay) in the ICU and hospital, and the SOFA score at day 7.  The only positive result was in the (prespecified) subset of patients with chronic arterial hypertension, with an interaction between study group and both doubling of creatinine and the need for renal-replacement therapy (which was said to be started for a blood urea nitrogen level of greater than 84 mg per deciliter or a creatining of greater than 5.65 among other criteria), with those outcomes being more frequent in the low target groups.  Inspection of Table 2 shows that, among patients without chronic hypertension, those outcomes were numerically higher in the high target groups.  But now we're torturing the data - this was not a trial testing targets in patients with chronic hypertension, so these findings are technically hypothesis generating.  But I suspect people will still want to change practice based on them, before the conduct of a trial limited to patients with chronic hypertension.

This trial is a welcome addition to our "evidence base".  It suggests that approaching normalization of the MAP ("normal" MAP for BP 120/80 = 93) does not appear to have a large beneficial effect.  (Among 350 patients with chronic hypertension, there were 20 "excess" cases requiring renal-replacement therapy in the low target group.)  The more interesting question for me is what is the lowest safe MAP that we should tolerate?  Surely there is some lower limit below which outcomes are worse because vital organs are inadequately perfused?  I am unconvinced, guidelines or not, that that lower limit is a MAP of 65.  Even though it is intuitively manifest that some degree of hypotension, left unchecked, will lead to organ failure and death, it remains possible that most of the hypotension we see with sepsis, especially after fluid administration and antibiotic provision, is an epiphenomenon of sepsis that will resolve as microbes die and endotoxin and cytokine levels decline.  We observe n=1 evidence of this (oh, the long maligned practice of empiricism which has become a lost art!) in those "natural experiments" which arise when the elderly patient with urosepsis has a DNR order and declines vasopressor administration.  Such patients very often survive without sequelae despite MAPs in the low 40s, oliguria, and mental status changes for several days.  I anxiously await reports of trials comparing a MAP target of 65 versus a target of 50-55 (corresponding to a BP of  around 80/40; to play it safe, it might be wise to exclude patients with chronic hypertension) - you can guess what my bet will be on the results of these trials.

1 comment:

  1. I fully agre with that. We are enough reading trials that deal with the epiphenomena, built on the false premise that we should normalize physiologic variables.

    After proper antibiotic administration, one can only expect small marginal results when comparing supportive strategies.