Web site designed and maintained by Chris Cairns  © SICS EBM Group 2004                                  

Up

 

Sepsis: low dose steroids improve outcome.

 

When all trials (n=16) combined: no overall effect.


In studies of long courses (³5 days) of low dose (£300 mg hydrocortisone or equivalent) steroids in patients with severe sepsis or septic shock: reduction in 28-day and hospital mortality in steroid treated patients.   No increase in adverse events in steroid treated patients.


Level of Evidence: 1+           (heavily weighted to one RCT (n = 300))

 

Citation/s: Corticosteroids for severe sepsis and septic shock: a systematic review and meta-analysis. BMJ 2004; 329: 480 - 488.

Lead author's name and fax: D Annane, Critical Care Department, Hopital Raymond Poincare, Paris. djillali.annane@rpc.ap-hop-paris.fr

 

Three-part Clinical Question: In patients with severe sepsis and septic shock does the use of steroids lead to an improvement in outcome?

 

Search Terms: sepsis, severe sepsis, septic shock, steroids, corticosteroids.

 

The Review:

 

Data Sources: Cochrane Library, Medline, Embase, LILACS database, non-English sources, checked reference lists, and when possible contacted authors to identify additional published or unpublished data.

 

Study Selection: Searched for randomised or quasi-randomised trials, with or without blinding, on severe sepsis or septic shock in adults or children.

 

Data Extraction: Methodological quality: generation of allocation sequence and allocation concealment, method of randomisation and loss to follow up.

Primary outcome: all-cause 28 day mortality.    Secondary outcomes: ICU mortality, hospital mortality, shock reversal (stable haemodynamic status 24 h after weaning from inotropes), and adverse events (GI haemorrhage, super-infections, hyperglycaemia and "others").

The studies were multiple independent reviews of individual reports. They were tested for heterogeneity.

 

The Evidence:

23 studies found in total - 16 included (excluded: mixed population no separate septic shock data; ARDS only; or only abstract - incomplete outcome data)

16 studies: pooling all results - heterogeneity - no difference in 28-day or hospital mortality.

5 studies: long duration (³ 5 days), and low dose (£ 300 mg hydrocortisone or equivalent) steroids - in steroid treated patients reduction in risk of 28-day (NNT 9, 95% CI 6 to 33) and hospital (NNT 10, 95% CI 6 to 54) mortality.  

 

In four studies reporting ICU mortality: reduction in steroid group, in four trials reporting shock reversal: increase in steroid treated group.  

 

No increase in rate of adverse events in those treated with low dose steroids.

 

Comments:

 

1) Do the methods allow accurate testing of the hypothesis? - YES - Methodologically sound systematic review.

2) Do the statistical tests correctly test the results to allow differentiation of statistically significant results? - YES - The correct tests were done.   No overall effect when all steroid trials included, heterogeneity of effects.

3) Are conclusions valid in light of the results? - YES

4) Did results get omitted, and why? - NO - clear reasons given for exclusion of certain studies.

5) Did they suggest areas of further research?  Yes - better define patients who may benefit from steroids - particularly in relation to response to ATCH - 2 studies have used this approach, but used different thresholds.     What they don't highlight is the need for one BIG simple RCT - note all 16 trails included total of only 2063 patients, 5 trails of long duration, low-dose steroids only 465 patients.  For the recommendation they are making - these are low numbers.

6) Did they make any recommendations based on the results and were they appropriate? - YES - only give steroids to patients with either absolute adrenal insufficiency (random cortisol < 414 nmol.l-1), or relative adrenal insufficiency to ACTH test (cortisol increment < 248 nmol.l-1). This paper should be read in conjunction with Minneci's meta-analysis, which came to different conclusions.

7) Is the study relevant to my clinical practice?  YES - summarises effects of trials to date of steroids in severe sepsis and septic shock.    However, weighted heavily to one trial (n=300) - contributes about 70% to data on long duration (³ 5 days), and low dose (£ 300 mg hydrocortisone or equivalent) steroid trials.

8) What level of evidence does this study represent?  1+

9) What grade of recommendation can I make on this result alone?  B

10) What grade of recommendation can I make when this study is considered along with other available evidence?  A

11) Should I change my practice because of these results?    YES - in patients with severe sepsis or septic shock long duration (³ 5 days), and low dose (£ 300 mg hydrocortisone or equivalent) steroid therapy use should be limited to patients with either absolute adrenal insufficiency (random cortisol < 414 nmol.l-1), or relative adrenal insufficiency to ACTH test (cortisol increment < 248 nmol.l-1).

12) Should I audit my current practice because of these results? YES - what proportion of patient fulfilling study criteria have an ACTH test done?; what proportion of those fulfilling criteria listed above receive intravenous corticosteroids?

 

Appraised by: Malcolm Daniel, Department of Anaesthesia, Glasgow Royal Infirmary; 24 September 2004

E-mail: md23s@udcf.gla.ac.uk 

 

Citation: EBM Critical Appraisals. Scottish Intensive Care Society EBM Group. Daniel M. 2004. : Annane D, et al. Corticosteroids for severe sepsis and septic shock: a systematic review and meta-analysis. BMJ 2004; 329: 480 - 488.

 

Reviewed for the SICS EBMG by CC & SJM. 

 

Kill or Update By: October 2007

 

Printer Friendly Version

 

© SICS EBMG 2004