The prediction of adverse outcomes following major non-cardiac surgery

The prediction of adverse outcomes following major non-cardiac surgery is complex. Clinical variables and risk factors, functional status, electrocardiography and non-invasive cardiac investigations can all be used to assess and stratify the risk of post-operative cardiac morbidity or mortality. Multiple factors can be combined into bed-side scoring systems. Increasingly, cardiac biomarkers such as b-type natriuretic peptide (BNP) have been shown to predict heart failure and mortality in non-surgical populations. In the studies in this thesis, I have investigated the incidence of peri-operative cardiac morbidity and mortality in patients undergoing major non-cardiac surgery and identified clinical variables that predicted adverse outcomes. I have tested the utility of BNP for prediction of cardiac complications. I have investigated the long-term survival of the patients in the cohort to identify predictors of reduced survival. I have examined the predictive value of the pre-operative 12-lead ECG for adverse outcomes. I have also studied the utility of a commonly used risk scoring system, the revised cardiac risk index (RCRI), for prediction of cardiac events. The study was a prospectively performed observational study of consecutive patients undergoing major surgery. The cohort consisted of patients undergoing aortic surgery (25.8%), lower limb bypass surgery (29.8%), amputation (25.2%) and laparotomy (20.0%). The patients underwent post-operative screening for myocardial infarction; consisting of serial ECG and troponin measurement. The end-points were major adverse cardiac event (MACE), defined as myocardial infarction or cardiac death and all-cause mortality. Long term follow-up was performed following discharge. Three hundred and forty-five patients were recruited to the trial. Forty-six patients (13.3%) suffered a peri-operative MACE and twenty-seven patients (7.8%) died in the post-operative period (six weeks). Independent predictors of peri-operative MACE were pre-operative anaemia, urgent surgery, a history of hypertension and age > 70 years. Pre-operative BNP was significantly higher in patients who subsequently went on to have a peri-operative MACE, compared with those who did not. An elevated BNP was an independent predictor of both MACE and peri-operative mortality on multivariate analysis. A low BNP was highly indicative of an uneventful post-operative period, with a negative predictive value of 96% for MACE and 95% for all-cause mortality. Traditional clinical markers of heart disease, such as past history of ischaemic heart disease, prior myocardial infarction, cerebro-vascular disease or history of cardiac failure provided no predictive utility for either MACE or mortality. The mortality rate at 1 year was 19.1%. The median follow-up period was 953 days (IQR 661-1216 days). Age > 70 years, diabetes, hypertension, renal impairment, a history of left ventricular failure, anaemia and urgent surgery were associated with reduced long-term survival. A BNP concentration of 87.5 pg/ml provided the best combined sensitivity and specificity for prediction of long-term mortality. Patients with an elevated BNP (>87.5 pg/ml) had a significantly reduced survival and BNP >87.5 pg/ml independently predicted reduced survival on Cox regression analysis. Urgent surgery and anaemia were also independent predictors of reduced long-term survival. An abnormal ECG was observed in 41% of patients recruited. An abnormal ECG was associated with an increased peri-operative MACE and mortality rate. Ventricular strain and prolonged QTc (>440ms) were ECG abnormalities that predicted MACE on multivariate analysis. Patients with an abnormal ECG, but no prior cardiac history, represent a high risk group that may benefit from optimisation. The studies in this thesis have identified that BNP, a simple pre-operative blood test, provides valuable information regarding the risk of both peri-operative morbidity and mortality, and long-term survival after major non-cardiac surgery. Improved risk stratification could allow targeted intervention and medical optimisation prior to surgery with the aim of modifying the risk of adverse outcomes