Gold 1995.

Study characteristics
Methods	Study design: randomised, open‐label controlled clinical trial Study duration: 30 months between October 1991 and February 1994 'Run‐in' period: none Number of study centre and location: 1 in the US
Participants	Total number of study participants: 251 Number of randomised participants: 248 Number lost to follow‐up: 0 Number withdrawn: 0 Number analysed: 248 Mean age: 65.8 years Gender: 160 men Inclusion criteria: people undergoing primary elective multivessel CABG for left main or multivessel coronary artery disease Exclusion criteria: inability to complete the neuropsychological tests (blindness, deafness, language difficulties), participation in other studies, and inability to return for follow‐up.
Interventions	Experimental: maintaining MAP during CPB at 80–100 mm Hg Comparison: maintaining MAP during CPB at 50–60 mm Hg Concomitant medications: CPB flow by body surface area and temperature were held constant, and vasoactive drugs were used to maintain MAP in the desired range. Anaesthesia was induced with thiopental (1–2 mg/kg), fentanyl (25 μg/kg), and pancuronium, and was maintained with a fentanyl bolus (1–5 μg/kg, to a total of 50–70 μg/kg), midazolam, or isoflurane (pre‐CPB and post‐CPB periods only). After sternotomy and pericardial incision, heparin was administered to maintain an activated clotting time > 480 seconds. After cannulation of the aorta and right atrium, non‐pulsatile CPB was instituted. Flow rates were set at 1.6 L/minute/m² during cooling and 2.4 L/minute/m² during warming. If the MAP increased above the target level and was unresponsive to fentanyl or midazolam, sodium nitroprusside infusion was administered. If the MAP fell below the target level, phenylephrine was used. If necessary, noradrenaline or metaraminol was added. Intraoperative ischaemia was managed by an identical algorithm in both groups. Excluded medications: none
Outcomes	Mortality, cardiac morbidity, neurological morbidity, deterioration in cognitive status, and deterioration in quality of life reported at 6 months. Cardiac complications were myocardial infarction, pulmonary oedema, adult respiratory distress syndrome, low flow state/cardiogenic shock, and cardiopulmonary arrest. Definite stroke was the principal neurological complication determined by the neurologist. Stroke included the new onset of a localised and persistent neurological deficit (e.g. paresis, plegia, aphasia, hemianopsia, cortical blindness). Deterioration on ≥ 3 cognitive tests was defined as a cognitive complication. For each test, the assessment was based on within‐patient change in test performance from preoperative baseline. Changes from preoperative to postoperative function that would be considered clinically important were determined a priori by a panel of experts. Deterioration in quality of life was defined as a decline of > 5 points on the Physical Component Summary score of the SF‐36.
Notes	Funding source: grant HL44719 from the National Institutes of Health, National Heart, Lung, and Blood Institute.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation based on a table of random numbers.
Allocation concealment (selection bias)	Unclear risk	Information not provided.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Information not provided.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Study cardiologist and neurologist, blinded to the intraoperative management, performed standardised examinations at 1, 2, and 7 days, and 6 months after the operation.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up among the randomised participants.
Selective reporting (reporting bias)	Unclear risk	Information not provided.
Other bias	Low risk	Review authors believed the study to be free of other sources of bias.