SURVEILLANCE FOLLOWING ENDOVASCULAR REPAIR OF ABDOMINAL AORTIC ANEURYSM

Introduction Endovascular Aneurysm Repair (EVAR) is the commonest form of repair of Abdominal Aortic Aneurysms (AAA). EVAR involves the placement of a stent-graft that seals into the arteries proximally and distally containing blood flow through the AAA. EVAR suffers from complications and treatment failures and patients are enrolled on a surveillance programme to identify these so corrective secondary intervention might be undertaken. Surveillance contributes significantly to the cost of treatment and is inefficient. Efficient surveillance requires adequate knowledge of the; nature and timing of complications, diagnostic merits of surveillance imaging, factors which limit patient compliance and ability to predict the risk of future complications. This work addresses deficiencies in these areas. Methods A database of a city wide vascular surgical service and surveillance imaging reports for 2008-15 were analysed under ethical approved. The nature / incidence of secondary interventions and degree of compliance with surveillance were reported using descriptive statistics and incidence. Association between patient factors and non-compliance were examined using adjusted odds ratios (AOR). Missing datapoints in imaging data were addressed using multiple imputation utilising chained equations. Surveillance imaging findings were assessed for association with subsequent secondary intervention using Kaplan-Myer plots, log rank test and AORs. AOR were calculated using multivariate models. A piecewise exponential model (PEM) was created to predict future risk of secondary intervention at different times following EVAR. This was internally validated using a bootstrapping technique and assessed using Hosmer–Lemeshow test and receiver operator characteristics (ROC). A prospective imaging study of Contrast Enhanced Ultrasound Scan (CEUS) and time-resolved Computer Tomography Angiography (tCTA) was undertaken, the diagnostic values of CEUS to diagnose a graft-related endoleak were obtained. Results A total of 2901 patient years of follow-up in 756 individuals were analysed. The intended purpose of the 178 secondary interventions was, 85(48%) maintaining distal perfusion and 93(52%) maintaining effective aneurysm treatment. The incidence of secondary interventions following EVAR varied at different stages of follow-up between approximately 35 and 60 interventions per 1000 patient years. Compliance was maintained above 95% until 10 years after EVAR. The factors most associated with compliance were; years after EVAR (AOR 0.62), calendar year EVAR performed (0.67), age at time (0.96) and previously undergone a secondary intervention (1.13). Evaluation of Colour Duplex Ultrasound Scan (CDUS) and Abdominal Radiography (AXR) findings demonstrated that reducing the number of collected findings did not lead to changes in the overall association with secondary intervention. Internal validation of the PEM model demonstrated an area under the curve of 0.72 (95% CI 0.68 – 0.76) on ROC analysis and Hosmer–Lemeshow test produced a median p-value of 0.51, demonstrating satisfactory discrimination and calibration. On blinded prospective study CEUS has a sensitivity of 0.56 (95% confidence interval 0.23 – 0.88) and specificity of 0.90 (0.78 – 1.00) to diagnose a graft related endoleak, with tCTA as the comparator standard. Conclusion This work demonstrates the changing epidemiology of secondary interventions after EVAR, that CDUS & AXR reporting can be simplified and those findings can be used to predict future risk of secondary interventions. Excellent compliance with EVAR surveillance is achievable and is associated with a small number of patient factors. Finally, CEUS is demonstrated to have poor graft related endoleak diagnostic values when compared to tCTA