Modeling the Long-term Cost-Effectiveness of Endovascular or Open Repair for Abdominal Aortic Aneurysm

Abstract

Abstracts Gregory L. Moneta, MD, Section Editor Atherosclerotic Plaque Composition and Occurrence of Restenosis After Carotid EndarterectomyHellings WE, Moll FL, De Vries JP, et al. JAMA 2008;299:547-54. Conclusion: Lipid-rich inflammatory plaques are associated with a reduced risk of restenosis after carotid endarterectomy.Summary: Both clinical and angiographic criteria have been used to try to determine the risk of restenosis after a vascular intervention. The authors of this study evaluated the composition of the atherosclerotic plaque at the intervention site in terms of plaque features that may be related to carotid restenosis after carotid endarterectomy. There were 500 patients prospectively followed up between April 1, 2002, and March 14, 2006. Patients were assessed for carotid artery restenosis as measured by duplex ultrasound imaging 1 year after intervention. A Ͼ50% stenosis was defined as a peak systolic velocity of Ͼ125 cm/s and a Ͼ70% stenosis was defined as a peak systolic velocity of Ͼ230 cm/s. Carotid restenosis after carotid endarterectomy was determined with predefined histologic plaque characteristics. These included macrophage and smooth muscle cell infiltration, collagen, calcification, interplaque hemorrhage, luminal thrombus, and lipid core size. Comparisons were determined using multivariate logistic regression analysis and were adjusted for clinical characteristics.At 1 year, 85 patients (17%) had developed Ն50% carotid restenosis, and 40 (8%) developed Ն70% restenosis. Histologic examination of the plaque revealed that increased macrophage infiltration (n ϭ 286) had a lower risk of Ն50% restenosis than plaques with no or minor macrophage infiltration (n ϭ 215; 11.5% vs 24.3%; adjusted odds ratio [OR], 0.43; 95% confidence interval [CI], 0.26-0.72). Patients with higher macrophage infiltration also had a lower risk of developing Ն70% restenosis (4.5% vs 12.6%; adjusted OR, 0.36; 95% CI, 0.17-0.74). The 177 patients whose plaque had a lipid core that was Ͼ40% of plaque volume also had a lower risk of Ն50% restenosis than the 94 patients with a lipid core size of Ͻ10% (11.3% vs 25.5%; adjusted OR, 0.40; 95% CI, 0.19-0.81). A large lipid core also had a lower risk of developing Ն70% restenosis (5.6% vs 14.9%; adjusted OR, 0.42; 95% CI, 0.17-1.04).Comment: Many of the findings of this study are unexpected. Plaque characteristics of inflammation and higher lipid content, commonly thought to be associated with more dangerous plaques were, in this study, associated with lower rates of restenosis after carotid endarterectomy. Their other findings also seem in opposition to previous studies or prevailing opinion. For example, an incidence of 21% of Ͼ50% restenosis with Dacron patch angioplasty at 1 year is higher than usually reported. Also, vein and Dacron patches are not generally regarded as having much difference in their ability to prevent restenosis after carotid endarterectomy, but this study found vein patches were more much effective in preventing restenosis than Dacron patches. The...

Accounting for Uncertainty in Health Economic Decision Models by Using Model Averaging

“…To produce a sample of size N from the posterior predictive distribution of expected incremental cost and effectiveness, averaged over models M 1 and M 2 , N p.M 1 |x/ Monte Carlo replicates from model M 1 were merged with N p.M 2 |x/ replicates from model M 2 . In the probabilistic sensitivity analysis, to be consistent with the published analysis (Epstein et al, 2008) under the base case model, log-normal probability distributions were assigned to all hazard ratios (parameter (a) and (b) in Section 2) and gamma distributions assigned to the event rates (parameter (c)). Means and variances correspond exactly to the estimates and confidence intervals that were presented in Section 2.…”

“…The EVAR1 trial (EVAR Trial Participants, 2005) compared endo-vascular aneurysm repair (EVAR) with repair by open surgery in patients with large AAAs. Following the trial, which had an average follow-up of 3 years, a long-term model was developed (Epstein et al, 2008) to assess the lifetime costs and benefits of EVAR compared with open repair for 74-year-old men. Following aneurysm repair, if patients survived at least 30 days after surgery, they were assumed to enter a long-term Markov model.…”

“…The transition probabilities between the states were informed by data from the EVAR1 trial, population life tables and expert judgement. Further details of the model are given by Epstein et al (2008). For example, the risk of death from CVD at any time for a patient receiving EVAR is the product of the mortality rate for CVD in the general population (from population data), the hazard ratio for CVD death among the trial population relative to the general population (from expert judgement) and the hazard ratio for EVAR compared with open repair (from the EVAR1 trial).…”

“….1/ This depends on a 'threshold' λ, the amount of money that a policy maker is willing to pay for 1 unit of benefit (such as 1 QALY). The cost-effectiveness results for 74-year old men were presented by Epstein et al (2008) for a plausible 'base case' followed by a series of eight alternative sets of reasonable model assumptions. The probability that EVAR was cost effective was substantially different from the base case under three alternative scenarios, defined by the following parameters.…”

Mid-term cost-effectiveness analysis of open and endovascular repair for ruptured abdominal aortic aneurysm

“…The choice of treatment for rAAA was based on surgeon's preference and aneurysm anatomy during the interval from commencement of the study until October 2009, beyond which the unit began randomization as part of the IMPROVE trial 11 . IMPROVE randomized patients with a clinical diagnosis of rAAA to either immediate transfer to theatre for open repair or computed tomography (CT) followed by EVAR if anatomically appropriate.…”