Clinical Rountable, Take 2: Steroids & Sepsis

[proman]

Advertisement - Members don't see this ad

We've got to do something to keep this forum alive:

The January 10 2008 issue of NEJM published an article on hydrocortison and septic shock. Briefly, this was a multicenter, RCT double-blinded placebo controlled. 251 patients received 50mg hydrocortisone IV and 248 received placebo q6 hours x 5 days then tapered. Primary outcome was 28 day mortality among patients who did not respond to a stim test. There was no difference in mortality, but the treated patients had faster resolution of their shock. There were more cases of superinfection in treated patients.

Is this consistent with your experiences? Is the benefit of shorter shock duration worth an increase in infection? What is the current practice of steroids in shock at your institution? And, perhaps most importantly, will your practice change because of this article?

 

[Hernandez]

We've got to do something to keep this forum alive:

The January 10 2008 issue of NEJM published an article on hydrocortison and septic shock. Briefly, this was a multicenter, RCT double-blinded placebo controlled. 251 patients received 50mg hydrocortisone IV and 248 received placebo q6 hours x 5 days then tapered. Primary outcome was 28 day mortality among patients who did not respond to a stim test. There was no difference in mortality, but the treated patients had faster resolution of their shock. There were more cases of superinfection in treated patients.

Some things that stuck out in my mind about this article.

"499 patients ......were enrolled from March 2002 to November 2005, 52 participating intensive care units (ICUs)"​

only 499 pts in 3.5 years in 52 ICUs? There really seems that there might be some selection bias here, or at the very least, difficulty enrolling pts since pts who qualified were automatically given steroids without enrollment in the trial.

"
A sample size of 800 patients (400 per group) was needed to achieve a statistical power of 80% to detect an absolute decrease in mortality of 10% from an existing death rate of 50%"​

so by it's own statistics it is underpowered to detect the a 10% difference when that was the difference in mortality in the Annane trial.

3rdly

"the studies involved dissimilar populations of patients. In the Annane study, the patients had higher SAPS II scores at baseline, and the entry requirement for systolic blood pressure was less than 90 mm Hg for more than 1 hour despite fluid and vasopressor therapy; there also was a much higher rate of death at 28 days in the placebo group (61%, as compared with 32% in our study). Second, enrollment in the Annane study was allowed only within 8 hours after fulfilling entry criteria, as compared with a 72-hour window in our study. Third, fludrocortisone was not given to patients in our study, since 200 mg of hydrocortisone should provide adequate mineralocorticoid activity"​

This study probably won't affect my use of steroids in septic shock.

 

[legrita]

Can I venture a reply here?

Our practice setting is sort of changing because of this. It has been recommended the latest surviving sepsis campaign (see below). BUT, we had 1 critical care who was not responding to pressors and fluid resus. We recommened it to the intensivist who gave the pt a 1 time dose. It didn't work and the pt died. According to the guideline, it's supposed to be a drip....so, hope it works next time.


Steroids
 Consider intravenous hydrocortisone for adult septic shock when hypotension responds poorly
to adequate fluid resuscitation and vasopressors (2C)
 ACTH stimulation test is not recommended to identify the subset of adults with septic shock
who should receive hydrocortisone (2B)
 Hydrocortisone is preferred to dexamethasone (2B)
 Fludrocortisone (50 g orally once a day) may be included if an alternative to hydrocortisone
is being used that lacks significant mineralocorticoid activity. Fludrocortisone if optional if
hydrocortisone is used (2C)
 Steroid therapy may be weaned once vasopressors are no longer required (2D)
● Hydrocortisone dose should be
300 mg/day (1A)
● Do not use corticosteroids to treat sepsis in the absence of shock unless the patient’s
endocrine or corticosteroid history warrants it (1D)

 

[bulgethetwine]

We've got to do something to keep this forum alive:

The January 10 2008 issue of NEJM published an article on hydrocortison and septic shock. Briefly, this was a multicenter, RCT double-blinded placebo controlled. 251 patients received 50mg hydrocortisone IV and 248 received placebo q6 hours x 5 days then tapered. Primary outcome was 28 day mortality among patients who did not respond to a stim test. There was no difference in mortality, but the treated patients had faster resolution of their shock. There were more cases of superinfection in treated patients.

Is this consistent with your experiences? Is the benefit of shorter shock duration worth an increase in infection? What is the current practice of steroids in shock at your institution? And, perhaps most importantly, will your practice change because of this article?

Fantastic!

I swear to God, I logged on today specifically to start a thread on this to keep things alive. I read this article (and the editorial) and promptly dusted off a copy of the Annane article for comparison:

Effect of treatment with low doses of hydrocortisone and fludrocortisone on mortality in patients with septic shock. JAMA 288: 862-871.

I think reviewing both articles simultaneously is helpful.

I think that my practice is going to stay the same -- I'm going to continue to use steroids in severe sepsis. But it sure has raised some interesting questions. Here are some things floating around in my head:

1) What specific advantage might fludrocortisone have had?
2) I think the patients in the Annane article were sicker.
3) How important is it to start the steroids ASAP? 8 hours in the Annane data, 72 in the Sprung. And, if we did an RCT in an ED setting, this time could get cut to almost zero. The effect might all be time dependent .... not such a 'far out' thought since, for example, if you give steroids prior to or simultaneous with antibiotics for meningitis in s. pneumo, the steroids have utility, but if you give it later, no advantage.
4) How important is etomidate, really, in affecting the adrenals? Some say a single dose. Maybe we shouldn't be using etomidate in suspected sepsis cases?
5) I think we need more end points! The Sprung article shows quicker recovery from the septic state but no incresae in mortality... and the best we've got to explain it is increased infection rates? Maybe. But I think there's more.
6) The Sprung study was surely underpowered. Only a 35% chance to uncover a 10% RR in death isn't powerful enough for me to change my practice.
7) How good IS the stim tests we use?


Crit care forum on the comeback, oh ah! Forum on the comeback, oh ah! Forum on the...

 

[Hernandez]

Fantastic!

4) How important is etomidate, really, in affecting the adrenals? Some say a single dose. Maybe we shouldn't be using etomidate in suspected sepsis cases?

This is something that is starting to emerge and should be studied. I'm questioning whether or not I should use it as a sedative for a RSI style intubation.

 

[bulgethetwine]

This is something that is starting to emerge and should be studied. I'm questioning whether or not I should use it as a sedative for a RSI style intubation.

I finally got around to critically examining the Annane article side by side with the Strung piece.

This was my review:

Review of Annane Trial:

-Randomized; blinded; similar groups, treated equally; results were analyzed in an intention to treat manner. Good study.
-Primary end point: death at 28 days: 10% RR in the non-responders; 6% RR for all patients.
-NNT: 7 and 8 respectively

Was it proof positive to give steroids? Still controversial. Sepsis campaign: Grade C recommendation – large scale randomized trails still needed to be done.

So, along (finally) comes the CORTICUS group (Sprung et al)

-Randomized, blinded, well-balanced, intention to treat analysis
-No difference in primary end point of death at 28 days.
-Secondary end point benefit: time until shock was reversed was shorter among patients who received hydrocortisone regardless of their responder/non-responder status.
-Also found a downside: superinfection

WHAT NOW? Show of hands – who’s going to give steroids TOMORROW?

POSITION: In my opinion, Sprung paper adds little to change our practice

Why won’t Sprung et al have a significant impact?

-Underpowered: 35% change of uncovering a 10% relative risk reduction – so 13 times out of 20, this study would fail to uncover a significant, positive result even it is really there – a Type II error!
-Dissimilar populations: Patients in the Annane study were sicker based on higher SAPS II scores, higher death rate in placebo group, all patients required mechanical ventilation
-Time to treatment: Annane: < 8 hours; Sprung: up to 72 hours!
-Antibiotics – retrospective pharmacy review assessing adequate coverage: > 90% for Annane vs. 72 – 78% for Sprung
-Steroid regimen: Hyrdocort + Fludro for Annane; Hydro only for Sprung (and this is weird, too, because although some sources claim this is okay because hydro provides adequate mineralo, it's not like there's actually a human study to prove it! Why didn't CORTICUS just use the same damn regimen?!?
-Steroid duration: -7 days (Annane) vs. 11 days (Sprung) - this might matter for the whole superinfection theory.
-Is superinfection really, truly a factor? Sprung et al is the only modern trial that showed an increase in superinfections – in contrast to the Annane trial, a metanalysis by a different group, and even the ARDS Network trials.
To decide if superinfection is really related to the steroids, a dedicated arm of a future trial should exist, not a makeshift “guesstimate” of whether each new infection is due to the steroids

Conclusion I:
We still await a randomized trial with relevant end points, adequate power, considered and appropriate inclusion criteria, and one that standardizes for the use of etomidate.

Conclusion II:
In the mean time, for your adequately fluid resuscitated, intubated, septic patients still requiring pressors to maintain SBP > 90mmhg after one hour, give hydrocortisone. This conclusion is actually supported by both parties. (I"m actually going to give fludrocortisone, too.)

Conclusion III: The cortisol stim test really has no real utility. Don't stop the steroids until you have a better test to prove that your patient isn't experiencing some degree of adrenal insufficiency.