Methodological aspects on carotid artery intervention

Abstract

In Sweden, approximately 25000 patients suffer a cerebral infarction (stroke) annually. Stroke ranks as the third-leading cause of death, and is the leading cause of functional impairment. Carotid artery stenosis accounts for approximately 15% of all strokes and almost all of these are theoretically preventable. There are two conceptually different invasive methods used to treat carotid stenosis: carotid artery stenting (CAS) and carotid endarterectomy (CEA). There are various aspects of these treatments that remain to be evaluated in order to minimize the risks of poor outcome, namely: any differences in short-term and long-term outcomes after CEA or CAS, the potential risks of CAS early after a previous transient ischaemic attack (TIA)/stroke, and the methods to indicate the need of shunting during CEA. Furthermore, since data from randomised controlled trials (RCTs) are often from specialized high volume centres with selected patients, it is important to evaluate population-based results after CAS and CEA.

The overall aim of this thesis was to investigate aspects of invasive treatment of carotid artery stenosis, in particular the outcome following endovascular treatment (CAS), and the efficacy of near-infrared spectroscopy in predicting selective shunt use during CEA. The outcome after endovascular treatment was investigated in three cohort studies (study I, II, and III). In study I, we analysed Swedish national short-term results of CAS and CEA and compared a high-volume centre with national results for CAS. In study II we evaluated the procedural risk of CAS, in relation to time from previous TIA /stroke. Study III assessed the long-term population-based outcomes after CAS as compared to CEA. In study IV, a multi-centre diagnostic study, we evaluated the ability of near-infrared spectroscopy (NIRS) to predict cerebral ischaemia during CEA.

Study I was a retrospective single-centre review of all consecutive patients treated with CAS at Södersjukhuset between 2004-2011. These results were compared to data from the Swedish national registry for vascular surgery (Swedvasc), where data from all patients treated with CAS or CEA during the same period was extracted. We found that the Swedish national results for CAS were inferior to CEA, but also that it was possible to achieve acceptable results with CAS in a high volume centre. During the study period 464 CAS procedures were performed in Sweden, with 208 of them performed at Södersjukhuset. The stroke and death rate at 30 days among patients in Sweden (Södersjukhuset excluded) was 7.4%. In Södersjukhuset, the stroke and death rate was 2.9%. The results for CAS on a national basis improved over time, the stroke and death rate during the first half of the period was 6.4%, compared to 3.6% in the second part of the study period, which was probably due to the learning curve for the procedure.

In study II, all CAS performed for symptomatic carotid stenosis between 2005-2014, were included in a retrospective analysis. Study II revealed that it was safe to perform CAS early (within 1 week), following a TIA/stroke. The stroke and death rate for patient treated within 7 days from neurologic symptom was 4.1%, compared to 6.3% between day 8-14, 4.8% between day 15-28, and 3.6% between days 29-180.

The long-term results after CAS in Sweden were examined in study III. In this study, all primary CAS procedures registered in Swedvasc were included (all reoperations were excluded). We compared 409 CAS patients to a control group treated with CEA. The control group was matched with respect to known confounding factors. Postoperative stroke was identified by cross matching the cohort with the In-patient registry. Median follow-up time was 4.1 years, and follow-up data was almost complete. By using Cox regression, we found a substantial (59%) increased risk for late(>30 days) stroke or death for patients treated with CAS compared to CEA. This increased risk was mainly due to an increased risk of ipsilateral stroke during long-term follow-up.

In study IV, we have shown that near-infrared spectroscopy has a high sensitivity and acceptable specificity to predict cerebral ischemia during CEA. When a cut-off point of 9% relative decrease in regional SO2 during carotid clamping was chosen, the sensitivity for detecting cerebral ischemia was 95%, and the specificity 81%. This was clearly superior to the conventional and widespread technique of measuring, “stump pressure”.

In conclusion, the national population-based results for CAS were inferior to CEA, both short- and long-term. The short-term results after CAS in Sweden have improved over time, and acceptable results can be achieved in high volume centres. If CAS is performed due to symptomatic stenosis it could be performed early after onset of neurologic symptoms. Near-infrared spectroscopy is a reliable method for selective shunting during CEA and superior to stump pressure.