• primary efficacy endpoint is all cause mortality [ Time Frame: 28 days ] [ Designated as safety issue: Yes ]
  
• primary toxicity endpoint is Serious Adverse Events, excluding sepsis-related events specified as secondary outcomes [ Time Frame: 28 days ] [ Designated as safety issue: Yes ]
  

• PIM2 [ Time Frame: entry ] [ Designated as safety issue: No ]
  
• PELOD [ Time Frame: daily to 28 days or PICU discharge ] [ Designated as safety issue: No ]
  
• ICU mortality [ Time Frame: 28 days ] [ Designated as safety issue: Yes ]
  
• time until shock reversal, defined as cessation of inotropic support for 24 hours [ Time Frame: 28 days ] [ Designated as safety issue: No ]
  
• time to resolution of multiorgan dysfunction [ Time Frame: 28 days ] [ Designated as safety issue: No ]
  
• time to resolution of base deficit [ Time Frame: 28 days ] [ Designated as safety issue: No ]
  
• time to resolution of lactate [ Time Frame: 28 days ] [ Designated as safety issue: No ]
  
• time to decision to discharge from ICU [ Time Frame: 28 days ] [ Designated as safety issue: No ]
  
• laboratory analysis of adrenal function [ Time Frame: 6 days and convalescence ] [ Designated as safety issue: No ]
  
• laboratory analysis of inflammatory parameters (defined in protocol) [ Time Frame: 6 days and convalescence ] [ Designated as safety issue: No ]
  
• laboratory analysis of coagulation parameters (defined in protocol) [ Time Frame: 6 days and convalescence ] [ Designated as safety issue: No ]
  

1. PURPOSE: The Need for a Paediatric Trial of Steroids in Sepsis - potential benefits and risks Numerous targets for new therapies in sepsis have been identified, none of which have been shown to have been of benefit in children. The results of adult studies cannot therefore be extrapolated directly to childhood disease. Corticosteroids alter the inflammatory balance in both beneficial and harmful ways in severe sepsis. Recent adult studies have demonstrated transient adrenal insufficiency is associated with adverse outcome and that corticosteroids increase survival in specific patient groups, and steroid replacement has become a standard of care. There is little uniformity in the approach to steroid replacement therapy amongst leading paediatric centres in the UK. Expert opinion has emphasised that guidance is interim while awaiting appropriate paediatric studies. Steroids are perceived as "safe" and "cheap" but should not be introduced into paediatric practice without further research. Sepsis in childhood differs in terms of mortality (around 10% overall in children vs in excess of 40% in adults), background immunity, co−morbidity, and causative organisms. Given the lower overall mortality in childhood sepsis, steroids have the potential to disrupt the inflammatory balance in children causing greater harm than benefit. It is not known which patients should be targeted for therapeutic intervention; what are the most appropriate endpoints; whether the length of steroid therapy can be shorter in children; or whether immunological rebound will occur.
2. DESIGN and METHODOLOGY:

This is an open randomised prospective pilot exploratory study of corticosteroid replacement therapy in three centres. Adrenal function measurements will be assessed on entry to the study. To investigate the inflammatory profile and the impact of corticosteroid replacement, blood will be taken for cytokine and coagulation protein analysis. This study will provide the pilot data necessary for the design of a definitive trial of corticosteroid replacement therapy with the identification of variables likely to improve our ability to stratify patients for intervention and the mechanistic characterisation of the modulatory effects of steroids on inflammation in children with severe sepsis. Enrolment will be undertaken in two stages (see flowsheet diagrams in protocol). Forty five eligible children will be randomly allocated to steroid replacement therapy for 2 days (n=30) or intensive investigation without intervention (n=15) in a 2:1 randomisation (stage 1); 45 subjects (stage 2) will then be randomly allocated to steroid replacement therapy for 5 days (n=30) or intensive investigation without intervention (n=15). Randomisation will the undertaken in accordance with a computer−generated list and will be stratified by age (<1 years; 1 year or more). Progression from stage 1 to stage 2 will follow an interim analysis by a Trial Monitoring Group to ensure safety. This escalating approach will provide safety data, information on length of therapy and an assessment of possible clinical and laboratory endpoints in addition to mortality, reducing the potential for adverse events in the pilot phase while providing data relevant to this population. A large excess of serious adverse events in stage 1 will result in study termination. After careful consideration by the investigators and during the peer review process, placebo will not be used in this study, which will inform a future large phase 3 randomised controlled trial.

RESEARCH PARTICIPANTS WILL RECEIVE THE FOLLOWING INTERVENTIONS THAT ARE NOT PART OF ROUTINE CLINICAL CARE (Please also refer to figures 1−4 in the protocol that we are unable to reproduce here): Children will be screened on admission to PICU. Entry into the study following consent involves a clinical test of endocrine function involving 2 blood tests. The list of procedures conducted in the study is as follows:

1. confirm eligibility requirements, assess pre−existing conditions and medical history, record weight, height, vital signs, data to inform clinical severity scores, complete infection assessment, clinically relevant laboratory investigations
2. corticotrophin stimulation test
3. multiple study samples (endocrine, cytokine and coagulation tests)
4. corticosteroid treatment if randomised to treatment group
5. follow−up in routine clinic