Invited commentary

Abstract

The positive aspects of this multicenter trial are that it is prospective, involves a large number of patients, and has had data from one of its phases analyzed by an independent core laboratory. The article is an important one, since the authors have previously identified poor fixation as the cause of most ruptures following endovascular repair with the AneuRx prosthesis.1Zarins C.K. White R.A. Fogarty T.J. Aneurysm rupture after endovascular repair using the AneuRx stent graft.J Vasc Surg. 2000; 31: 960-970Abstract Full Text Full Text PDF PubMed Scopus (209) Google Scholar It is unfortunate, therefore, that migration at modular connections and between the prosthetic limbs and native iliac arteries was not addressed. Since the article reports clinically significant events almost entirely on the basis of imaging findings, it is regrettable that details of imaging and methodology of measurement are not included in the text. It is well known that parallax error in plain radiographs makes detection of endograft migration by reference to bony landmarks on images obtained at different times an unreliable method. Similarly, in the absence of any information on the thickness of CT slices, the sensitivity of this modality in detecting migration is called into question. The criteria by which the core laboratory and clinical sites assigned the dichotomous endpoint of migration or no migration are missing, together with information on the intraobserver and interobserver error for measuring migration. Migration is defined by the authors as “any postimplantation movement or displacement of the stent graft.” If this statement is taken at face value, all 1025 patients categorized as having no migration would have had to have follow-up imaging with the endograft assigned the identical distance from the reference point as in the postdeployment images, accurate to the last millimeter. Information is also lacking on imaging data coming from participating clinical sites and not reviewed by the core laboratory. If these data were not obtained under any defined protocol and were produced by observers with variable skill levels, migration may have been underreported. Although the stated objective of the article is to identify factors that may predispose to stent-graft migration, the relatively high incidence of migration itself tends to overshadow the culprit factors of low initial deployment of the device below the renal arteries and a short proximal fixation length. The incidence of migration documented in the article is a cause for some concern. Two of the clinical sites had migration rates of 30% (Fig 4), while Kaplan-Meier analysis demonstrates an increasing incidence with time (18.8% at 3 years, Fig 2). Moreover, it should be noted that the mean follow-up from implantation is limited to 23 months, which is just short of a marked increase in events, and that 70% of patients did not reach the 3-year follow-up time point. Since the cut off date for this study was August 2001, surgeons and patients will await with interest the authors' next report.