The optimal timing of surgery following a stroke or transient ischemic attack (TIA) must balance the risks of the procedure against the benefit of preventing a recurrent stroke. In the past the preferred approach was to allow a waiting period of at least 2–6 weeks after symptom onset prior to considering operative intervention, based on concerns of higher perioperative complications and risk of hemorrhagic conversion [6]. This concern was prompted by multiple early studies which showed higher risks with early operation [7–9], and was highlighted by Giordano et al. in 1985, who noted an 18.5 % postoperative stroke rate in patients having surgery at less than 5 weeks and therefore recommended a 5 week waiting period [10]. These recommendations were based on studies during a time when the availability of CT and MRI imaging was limited and intensive care with aggressive blood pressure control was less common.

More contemporary studies have not supported these prior concerns. Numerous recent series have demonstrated equivalent or improved outcomes with earlier intervention for symptomatic carotid disease when compared to delayed repair [11–18]. Furthermore, it is now recognized that there is a significant risk of recurrent stroke without intervention, which ranges from 5 to 10 % at 7 days and 15–20 % at 30 days [19–21]. This has led to multiple guidelines recommending surgical treatment of symptomatic carotid stenosis within 14 days [22–25]. However, there remains controversy concerning the optimal time for intervention, especially in particular subsets of patients.

Although there are many studies evaluating the broad question of optimal timing for carotid intervention after stroke, the degree of heterogeneity and inconsistency between studies often makes comparison difficult. Definitions of early surgery, hyperacute surgery, degree of stroke, and methods of patient assessment sometimes differ widely between studies. In addition many of these studies were conducted – at least in part – before the availability of what we would now consider optimal medical management.

The large randomized multicenter trials North American Symptomatic Carotid Endarterectomy Trial (NASCET) and European Carotid Surgery Trial (ECST) provided the most robust patient cohort of symptomatic patients undergoing CEA [3, 5]. These studies demonstrated benefit for intervention in symptomatic patients with significant associated carotid stenosis and defined the important role of endarterectomy in management of these patients [3, 5]. Of note, these studies excluded patients with severe stroke and carotid occlusion, and the results are applicable only to patients with non-disabling stroke. A pooled analysis of data from these studies was conducted by Rothwell et al. in 2004 to evaluate the impact of timing on patient outcomes [18]. This study found that the greatest benefit to surgery was in patients who were randomized within two weeks of their ischemic event. The number needed to treat to prevent one stroke in 5 years increased from 5 patients in the <2 week group to 125 patients in the >12 week group. Absolute risk reduction from surgery in 5 year cumulative risk (ipsilateral stroke at 5 years and any stroke or death within 30 days of surgery) dropped from 18.5 % at <2 weeks to 9.8 % at 2–4 weeks, 5.5 % at 4–12 weeks, and 0.8 % at >12 weeks when considering symptomatic patients with >50 % carotid stenosis. This declining risk reduction with operative delay was even more significant when only patients with >70 % stenosis were considered, with drop from 30.2–17.6 % to 11.4–8.9 % at these same time intervals, respectively. For patients with 50–69 % stenosis the benefit of endarterectomy was erased after the 12 week mark. A separate NASCET and ECST subset analysis noted that women may have an even more striking decline in benefit from CEA with increasing time from symptom onset [26]. Based on these data, multiple consensus guidelines recommend intervention within the 2 week window following symptom onset [23, 27].

The goal of carotid surgery is to decrease the risk of stroke. Risk of recurrent stroke is highest in the first 7–14 days after the index neurologic event [20], and the risk of recurrent stroke while on medical therapy falls significantly over the following year [3, 5]. As such, moves toward earlier intervention should maximize benefit from stroke reduction, assuming that perioperative event rates are equivalent. Indeed, there is evidence that waiting longer after symptom onset may lead to additional strokes that could have been prevented with endarterectomy [28–30]. Rantner et al. noted that 11.8 % of patients waiting for a delayed CEA had a secondary stroke or carotid occlusion during the 4 week waiting period, with most occurring at the 3–4 week mark [28]. Analysis of the medical arm of the NASCET trial demonstrated a 90 day risk of recurrent neurologic event in 20.1 % of patients that presented with TIA, and 2.3 % of patients that presented with hemispheric stroke [29]. This study also noted a 5.5 % risk in the first 2 days alone for the TIA group. Ois et al. reported even higher risk of early stroke or TIA recurrence, with 27.6 % of patients presenting with mild stroke or TIA having recurrent neurological event, and 20.9 % occurring in the first 72 h [30]. Both NASCET and ECST subgroup analyses demonstrated the greatest benefit in regards to recurrent stroke in patients who underwent surgery within 7 days of the neurologic event [18, 31].

The concerns for increased perioperative risk in the early period after stroke or TIA must also be considered to evaluate the risk/benefit profile with early intervention in symptomatic carotid disease. Pooled analysis of the NASCET and ECST data demonstrated a 30 day stroke and death rate of 7 %[26], and published guidelines cite benefit to CEA when a surgeon’s perioperative risk is <6 % [32]. Numerous recent studies have evaluated this and generally shown similar risk profiles at different times of intervention [12, 13, 15, 17, 33, 34]. Rerkasem and Rothwell performed a systematic review in 2009 to evaluate operative risk based on the timing of surgery, and found no significant difference between early or later endarterectomy in patients with stable neurologic status after a non-disabling stroke or TIA [17]. Concerns about causing hemorrhagic conversion also appear to be largely unfounded in patients with mild stroke or TIA, with a Cochrane review demonstrating a 0.2 % incidence of intracranial hemorrhage in this patient subset [33]. On the other hand, patients with major stroke or stroke in evolution appear to be at substantially higher risk [17]. Ballotta et al. performed a prospective randomized study comparing early to late endarterectomy, which demonstrated no difference in perioperative stroke, survival, or stroke-free survival at 3 years [13]. Paty et al. [15] retrospectively looked at perioperative stroke risk at 1 week time intervals after CEA, and found no significant differences. Permanent deficit developing perioperatively occurred in 2.8 % of patients operated on at <1 week, 3.4 % at 1–2 weeks, 3.4 % at 2–3 weeks, and 2.6 % at 3–4 weeks. The only factor that was associated with post-operative permanent neurologic deficit was the size of the lesion on preoperative imaging. Annambhotla et al. [34] performed a retrospective analysis and found no difference between early (<30 days) versus late (>30 days) endarterectomy in regards to 30 day mortality, stroke, or MI. Ballotta et al. [12] performed a prospective study evaluating 102 patients who had CEA within 2 weeks of minor stroke or TIA. They did not have any perioperative deaths, strokes, or episodes of cerebral bleeding in their series, although 2.9 % of their patients did suffer transient neurologic deficits perioperatively. Cerebral infarcts on preoperative imaging studies did not affect the risk of subsequent neurologic impairment.