An analysis of sensitivity, specificity, positive predictive value and negative predictive value of left ventricular ejection fraction in patients undergoing abdominal aortic aneurysm repair

Abstract

Objective: Preoperative risk assessment is important in patients undergoing major vascular surgery in the current era of evidence based medicine. Identification of individuals at increased risk may allow important alterations in perioperative management. There has always been a debate as to which test is best in risk stratification. This study was performed to assess the sensitivity, specificity, positive and negative predictive value (PPV and NPV) of left ventricular ejection fraction (LVEF) to predict cardiac complications and 30-day mortality in patients undergoing abdominal aortic aneurysm (AAA) repair. Methods: The study was designed as a retrospective analysis of prospectively collected data. List of patients who underwent elective open AAA repair between October 2006 and December 2009 was obtained from departmental database. All patients who underwent assessment of LVEF by means of multigated acquisition (MUGA) scan were identified and included in the study. LVEF of 40% or more was labelled as normal as per protocol and was used as cut-off for inter-group analysis. Perioperative parameters and outcomes were recorded for all patients. Statistical analysis was performed using SPSS v16.0. Results: Seventy-three patients [62 male; median age: 75 years (IQR: 71-79)] were included in the study. Median hospital stay was nine days (IQR: 7-12). LVEF was <40% in 37% (n=27) and >40% in 63% (n=46) of patients. Postoperative cardiac complications were observed in 6.8% (n=5) patients. Thirty-day mortality was 9.5% (n=7). Sensitivity, specificity, PPV and NPV of LVEF for prediction of cardiac complications were 40%, 63%, 28% and 93%, respectively. Similar results were seen for prediction of 30-day mortality (sensitivity 43%, specificity 68%, PPV 14%, NPV 90%). Inter-group analysis: there was no statistically significant difference between the two groups for basic demographics, co-morbidities and preoperative medications. Also no statistically significant difference was observed in the length of hospital stay (P=0.249). Conclusions: LVEF is not an appropriate indicator of cardiac complications or 30-day mortality in patients undergoing AAA repair. Further studies are required to assess the role of surrogate markers like cardiopulmonary exercise testing in addition to LVEF for prediction of morbidity and mortality in these patients.