Giomarelli 2003.

Study characteristics
Methods	Prospective, randomised, double‐blind, placebo‐controlled clinical study Study period: between January 2002 and April 2002; accepted for publication 24 January 2003 Number of study centres and locations: single‐centre, Institute of Thoracic and Cardiovascular Surgery, University of Siena, Siena, Italy
Participants	Twenty patients undergoing CABG received either preoperative steroids (n = 10, protocol group) or no steroids (n = 10, control group). Mean age: 64 years Sex (female/male ratio): 35% No high‐risk population Inclusion criteria: not specified Exclusion criteria: urgency/emergency surgery; previous heart surgery; valve or combined CABG and valve surgery; LVEF < 0.35; diabetics on insulin therapy; active gastropathic disorder; COPD on therapy; preoperative use of steroids and contraindications to steroid administration; Cleveland Clinic score of 4 or higher
Interventions	Intervention group: steroid (1 g IV MP preoperatively and 125 mg at the end of CPB; in the ICU, 4 additional 125 mg doses were given every 6 hours) Control group: no steroids (similar volumes of isotonic sodium chloride solution at the same time)
Outcomes	TNF‐α, IL‐6, IL‐8, IL‐10, haemodynamic and pulmonary measurements (AaDo2, RI, shunt [Qs/Qt], dead space [Vd/Vt], a‐vo2D, OER, and oxygen delivery), heart protection (MAP, SVR, pulmonary artery pressure, PVR, CVP, cardiac index, CK and CK‐MB), and clinical outcomes (blood glucose level, extubation time, ICU or hospital LOS, postoperative major complication as adverse events, infection) Samples were measured at the following intervals: T1, 15 minutes after intubation; T2, 5 minutes after aortic cross‐clamp release; T3, 10 minutes after CPB; T4, 3 hours after CPB; T5, 12 hours after CPB; T6, 24 hours after CPB; and T7, 4 days after operation.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants were randomised according to a computer‐generated sequence. Participants were similar with regard to preoperative and intraoperative data.
Allocation concealment (selection bias)	Low risk	An anaesthesia nurse performed the randomisation and prepared the syringes of blinded solution that were administered by the anaesthesiologist managing the case.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind study: all caregivers were blinded to treatment group.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	All physicians and nursing staff caring for the participants perioperatively were unaware of the treatment groups.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the study (uneventful in the two groups); no trial group changes, no withdrawals, and no losses to follow‐up were reported, but no intention‐to‐treat analysis. Data from all the participants were included in the final analysis.
Selective reporting (reporting bias)	Unclear risk	Clinical data were not extensively reported as the other laboratory outcomes.
Other bias	Low risk	No funding was disclosed, and no other source for risk of bias was identified.