Hydrocortisone and Fludrocortisone in Septic Shock: A Patient Already Known to Be Suppressed
A single patient on chronic prednisone, now in septic shock with a cortisol of 9. Two trials, two regimens, and a mechanistic argument for the combination that turns out — on inspection — not to be pharmacologically true.
T.O. drove a city bus route for thirty-one years before his COPD forced him to retire five years early, and has been on a low daily dose of oral prednisone for chronic maintenance ever since — 5 milligrams, unchanged, for over two years, alongside a long-acting bronchodilator inhaler he still uses twice a day. His baseline function has been reasonably preserved despite the COPD; he still lives independently and, until this week, walked his dog around the block most mornings without needing supplemental oxygen. His daughter brought him in yesterday after two days of fever and confusion; a urine culture and imaging confirmed a complicated urinary tract infection with an obstructing kidney stone and early perinephric abscess, a source that had already progressed to septic shock by the time he reached the emergency department, and tonight he remains hypotensive on a moderate dose of norepinephrine despite adequate fluid resuscitation.
His random cortisol, drawn on admission, came back at 9 mcg/dL — lower than would be expected for someone under this degree of physiologic stress, and consistent with what years of daily exogenous prednisone would be expected to do to his own adrenal output. That number is the crux of tonight's argument: the two major trials behind adjunctive corticosteroids in septic shock, ADRENAL and APROCCHSS, were both built around genuine uncertainty about whether an average septic patient's adrenal axis is meaningfully suppressed. T.O. isn't that average patient — his suppression is already a documented, ongoing fact of his daily medical history, not a hypothesis the team needs biochemistry to test. He has also needed two short prednisone bursts for COPD exacerbations in the past year, on top of his daily maintenance dose, which the team notes only reinforces how chronically suppressed his own adrenal axis likely already is. What makes the case awkward is that the certainty runs in the wrong direction to help. Knowing for a fact that he is suppressed tells the team nothing about which regimen's outcome data transfers to him, because neither trial selected for patients whose suppression was already documented — ADRENAL and APROCCHSS both recruited on clinical shock, not on a known steroid history. He is the one kind of patient about whom the two trials are least informative, precisely because he is the one kind of patient about whom the diagnosis is most certain.
A patient the equipoise assumption may not fit
I'd start the full APROCCHSS regimen now — hydrocortisone plus fludrocortisone. That trial showed a real 90-day mortality benefit with the combination, and he's a more definite case for replacement than the average patient APROCCHSS enrolled under genuine uncertainty — he's on chronic exogenous steroids with a low cortisol to confirm it. If the combination helped a population where suppression was uncertain, it should help him more, not less.
I'd be more cautious extrapolating from APROCCHSS's mortality number alone. ADRENAL — a larger, more recent trial testing hydrocortisone by itself — found no mortality benefit, only faster shock resolution. His chronic steroid use is a real biological fact, but knowing he's suppressed doesn't by itself prove that treating for it changes his outcome any more than it did in ADRENAL's population. I'd give stress-dose hydrocortisone alone — physiologic replacement for his known suppression — without assuming the added fludrocortisone's benefit transfers.
You're right that his suppression is more certain than the average trial patient's — I'm just not convinced certainty about the diagnosis is the same thing as certainty about which regimen's outcome data applies to him.
I want to stop one argument before it gets into the orders, because I can hear it forming: that his chronic prednisone has left him with a mineralocorticoid gap that fludrocortisone should fill. That isn't true. Prednisone is not mineralocorticoid-free — its mineralocorticoid potency runs around 0.8 that of hydrocortisone per milligram, and dexamethasone is the agent with essentially none. More to the point, hydrocortisone at 50mg every six hours is 200mg a day, and roughly 20mg of hydrocortisone already approximates the mineralocorticoid effect of 0.1mg of fludrocortisone. Whatever else stress-dose hydrocortisone leaves him short of, mineralocorticoid activity is not it.
Which doesn't settle the question against you — it just moves it. If we add fludrocortisone, the reason has to be that APROCCHSS tested those two drugs together as a single regimen and that combination is what produced the 90-day mortality benefit, while ADRENAL tested hydrocortisone alone and found none. That's a defensible reason to reproduce a trial's regimen intact rather than unbundle it. It is a different argument from the one being made, and the distinction matters more than it looks: a mechanism that sounds right and isn't will outlive this admission and get applied to the next patient.
Agreed: start hydrocortisone 50mg IV every six hours plus fludrocortisone 50mcg enterally daily, with his home prednisone coverage folded into the hydrocortisone dosing rather than given as a separate additional dose.
Not agreed:
How long to continue the regimen past shock resolution. His own baseline chronic exogenous suppression complicates the usual rule of tapering shortly after vasopressors stop — the critical care physician wants to follow the standard septic-shock taper timeline regardless; the pulmonologist is worried a standard-speed taper would leave him under-replaced relative to his own established chronic need, and wants an endocrinology consult before any taper begins rather than defaulting to the usual protocol.