Introduction: The research undertaken in this thesis was to inform OPTIMISE, a randomised controlled trial of goal directed haemodynamic therapy (GDHT) versus usual care in high-risk patients undergoing gastrointestinal surgery. The trial involved a complex intervention of cardiac output monitored administration of fluids and inotropic drugs during the perioperative period. Uncertainty exists regarding: 1. Whether the choice of fluid therapy could have influenced the outcome of the trial. 6% hydroxyl-ethyl starch (HES) has been associated with risk of death and acute kidney injury (AKI) in critically ill patients. 2. Whether he availability or provision of critical care beds is associated with improved surgical outcome and thus could have influenced the outcome of the trial. The trial intervention has traditionally been administered in a critical care setting, and this may have a bearing on outcome. 3. The trial intervention itself could have been associated with increased cardiac complications. Concerns remain regarding the administration of inotropic agents outwith traditional indications. Methods: 1. A meta-analysis was undertaken comparing perioperative use of 6% HES solutions to any comparator. 2. Surgical activity, population demographics and critical care provision in the UK were examined using large administrative databases. 3. A UK-wide cohort of noncardiac high-risk surgical patients admitted to intensive care was generated by combining data held by the Scottish Intensive Care Audit Group (SICSAG) and the Intensive Care National Research and Audit Centre (ICNARC) for the calendar year 2009. 4. Using this data, advanced statistical modelling techniques were used to test the association between critical care bed provision and outcome after high-risk surgery. 5. Measurement of postoperative 5th generation highly sensitive troponin (HST) release was undertaken in a subgroup of trial participants, in order to determine if the intervention was associated with increased myocardial necrosis. Logistic regression was undertaken to test if preoperative measurement of HST was associated with risk of death or major adverse cardiac events (MACE). Results: The principal findings of this thesis were: 1. In a meta-analysis of 1567 patients from 19 clinical trials comparing perioperative administration of 6% HES solutions versus any comparator no difference was observed in 30-day mortality arms (p=0.91, 12=0%; FEM: RD 0.00, 95% CI -0.02, 0.02) or AKI (p=0.62,12=0%; FEM: RD -0.01, 95% CI -0.04, 0.02) was observed. 2. Significant variation exists in ICU bed provision within the UK. 3. In an epidemiological study of 16 147 patients admitted to ICU following surgery in the UK, significant variation in acute hospital mortality was observed (OR 1.42; 95% CI: 1.29, 1.62). This did not appear to be accounted for by severity of illness, other patient-level factors or ICU bed provision. 4. Using HST we were unable to detect any difference in myocardial injury or infarction between GDHT and usual care groups. Preoperative HST measurement did not predict those at risk of perioperative death or MACE. Conclusion: Use of 6% HES in the trial intervention was unlikely to have affected trial outcome. Significant regional variation exists in outcome after surgery in the UK, which cannot be account for by patient level-factors or ICU bed provision. The trial intervention in OPTIMISE was unlikely to have caused increased incidence of myocardial infarction or necrosis. In this study preoperative measurement of 5th generation HST did not appear to predict those at risk of death at 30 or 180 days or MACE.