Prior ischemic stroke is a strong risk factor for perioperative mortality and morbidity, including recurrent stroke and other major adverse cardiovascular events. These risks are highest in the months after stroke and decline over time. Increasing the interval between stroke and surgery may decrease the risk of perioperative complications. The benefits of delay must be weighed against the risks of postponing surgery. In this focused review, we examine 5 major studies of the timing of surgery after ischemic stroke. On the basis of this evidence, we provide a framework to guide the appropriate scheduling of surgery after stroke.
KEY POINTS
- Prior stroke is a strong risk factor for perioperative morbidity and mortality.
- Perioperative risk increases rapidly in the days after a stroke and declines over the course of months.
- Appropriate timing of surgery after stroke is a key strategy to mitigate perioperative risks.
Surgical stress induces widespread hemodynamic, endocrine, and inflammatory responses that may compromise cerebral and cardiovascular hemostasis in vulnerable patients.1 Cerebral blood flow autoregulation is impaired after stroke, which may magnify physiological insults.2 Perioperative fluctuations in blood pressure, cardiac output, and vascular tone as well as interruption of antithrombotic therapy place these patients at risk for recurrent stroke and other major adverse cardiovascular events (MACE).
Prior stroke has been consistently demonstrated to be a strong independent risk factor for perioperative stroke, MACE, morbidity, and mortality.3–6 Optimizing timing of elective surgery after stroke is critical in minimizing the risk of perioperative events while also avoiding potentially harmful surgical delays. However, evidence to guide optimal timing remains limited and consensus guidelines have evolved to reflect emerging research.
A seminal Danish cohort study published in 2014 explored how the risk of perioperative MACE varied with increasing time interval after ischemic stroke.7 This led the American College of Surgeons (ACS)8 and the Society for Neuroscience in Anesthesiology and Critical Care (SNACC)9 to recommend delaying elective procedures for 9 months after stroke. However, recent evidence has challenged the necessity of such prolonged delays. A 2022 analysis of Medicare beneficiaries found that risks of postoperative stroke and mortality stabilized ∼90 days after stroke.10 This led the American Heart Association to recommend elective noncardiac surgery be delayed at least 3 months after stroke or transient ischemic attack (TIA).11
Recommending at least 3 months between stroke and surgery does not imply that 3 months will minimize risk in all situations. Choosing an appropriate interval is a nuanced decision that must balance the stroke-related risks of early intervention against the risks from delayed treatment of surgical pathology. This focused review attempts to reconcile available literature and offer guidance for scheduling surgery for patients with recent stroke.
CURRENT EVIDENCE
Most studies examining risks of surgery after stroke include multiple logistic regression analyses, which report the adjusted odds ratio (aOR) and CI of adverse outcomes among patients with prior stroke grouped into discrete time intervals (Table 1). The reference groups are usually patients without prior stroke (or no stroke within a designated multiyear look-back interval). Several studies use restricted cubic splines15 to model change in risk over time in a continuous fashion. Since patients without prior stroke cannot be assigned a time value, the reference for these analyses is usually set as the median time between stroke and surgery in the cohort.
| Lead author | Population, groups and group size | Covariates in statistical models | Key outcomes assessed | Key findings |
|---|---|---|---|---|
| Jørgensen7 | All elective noncardiac, nonintracranial surgeries in Danish patients (≥20 y) from 2005 to 2011 grouped as follows: 1. No prior CVA (n=474046) 2. CVA <3 mos prior (n=862) 3. CVA 3 – <6 mos prior (n=469) 4. CVA 6 – <12 mos prior (n=898) 5. CVA ≥12 mos prior (n=4908) |
Sex, age, BMI, comorbidities, pharmacotherapy, type of surgery, surgical risk. | 30-day all-cause mortality, MACE (composite of nonfatal acute MI, nonfatal ischemic stroke, and cardiovascular death), and recurrent ischemic stroke. | Adjusted ORs were higher in the 3-<6 mos group compared with ≥12 mos group for 30-day MACE (4.85 vs. 2.47) and recurrent stroke (24.02 vs. 8.17). |
| Christiansen12 | All emergency noncardiac, nonintracranial surgeries in Danish patients (≥20 y) from 2005 to 2011 grouped as follows: 1. No prior CVA (n=135689) 2. CVA <3 mos prior (n=2289) 3. CVA 3-9 mos prior (n=1090) 4. CVA >9 mos prior (n=4117) |
Sex, age, BMI, comorbidities, pharmacotherapy, type of surgery, surgical risk. | 30-day all-cause mortality, MACE (composite of nonfatal acute MI, nonfatal ischemic stroke, and cardiovascular death), and recurrent ischemic stroke. | 30-day MACE was lower for surgeries within 3 d of CVA versus surgeries done 4-14 d after CVA (21.4% vs. 28.8%, P=0.029 among 323 PSM pairs). |
| Glance10 | US Medicare patients (≥66 y) admitted for elective noncardiac, non-neurologic surgeries from 2013 to 2018 grouped as follows: 1. No CVA in prior 720 d (n=5787506) 2. CVA ≤30 d prior (n=1517) 3. CVA 31-60 d prior (n=2356) 4. CVA 61-90 d prior (n=2612) 5. CVA 91-180 d prior (n=7611) 6. CVA 181-360 d prior (n=14675) 7. CVA 361-720 d prior (n=25262) |
Sex, age, race and ethnicity, dual-eligibility for Medicaid, community dwelling versus nursing home before surgery, comorbidities, surgical procedure class, and year of admission. | Primary: AIS during index surgical admission or re-hospitalized within 30 d of surgical admission with primary diagnosis of AIS. Secondary: 30-day all-cause mortality, composite of AIS or death within 30 d. |
Risk leveled off at different intervals depending on surgical risk and outcome. The adjusted ORs for recurrent stroke for the 61-90 d, 91-180 d, and 361-720 d groups were 5.01, 5.83, and 4.02, respectively, for the complete cohort. |
| Chalitsios13 | English adults (≥18 y) admitted for elective or emergent NHS-funded noncardiac, non-neurologic, surgery between 2007 and 2018. The exposure was prior events (stroke, TIA, MI, or unstable angina) grouped as follows: 1. Event 0-2 mos prior (n=97902) 2. Event 3-6 mos prior (n=55505) 3. Event 7-12 mos prior (n=71978) 4. Event 12-24 mos prior (n=115987) 5. Event >24 mos prior (n=535498) |
Sex, age, comorbidities, and index of multiple deprivation. Subgroup analyses included surgical invasiveness (minor, moderate, major). |
Primary: 30-day all-cause mortality. Secondary: Mortality at 60, 90, and 365 d. |
Mortality was higher in patients with surgeries 3 to 6 mos after an event vs. 7-12 mos after an event. |
| Luney14 | Same as Chalitsios13 | Same as Chalitsios13 | ACS, acute MI, or CVA within 1 y, prolonged length of stay, 30-day and 1-year unplanned readmission. | For elective surgeries, CVA risks remained elevated for >12 mos after both ischemic stroke and TIA. |
Timing of Elective Surgery After Stroke: Jørgensen et al. (Denmark)7
This 2014 study examined how the time interval between ischemic stroke and elective surgery affected the risk of 30-day MACE and mortality using Danish national registry data for the entire adult population. Cardiac, neurological, and aortic arch surgeries were excluded, as were tracheostomies and gastrostomies. The reference group was patients without stroke in the prior 5 years. Surgical risk was based on European Society of Cardiology guidelines, with high-risk surgeries comprising 4.5% of the stroke group and 1.2% of the reference group. The mean age of the stroke group (69.7±12.1 y) was significantly higher than the reference group (53.7±17.0 y).
Among 862 patients who had surgery <3 months after ischemic stroke, the authors observed high rates of 30-day MACE (aOR 14.23, 95% CI: 11.61-17.45), mortality (aOR 3.07, 95% CI: 2.3-4.09), and recurrent stroke (aOR 67.6, 95% CI: 52.27-87.42). Recurrent stroke accounted for more than half of the composite MACE events. The subgroup of patients having high-risk surgery showed lower odds of MACE (aOR 2.97, 95% CI: 0.98-9.01), consistent with an elevated baseline risk in that population.
Among patients with prior stroke, the risk of MACE was lower in patients taking statins and those on antithrombotic therapy. There was a stepwise decline in risk with increasing intervals, with statistically significant differences in the risk of MACE for the 3- to <6-month group (aOR 4.85, 95% CI: 3.32-7.08) compared with patients whose stroke occurred ≥12 months before surgery (aOR 2.47, 95% CI: 2.07-2.95). Cubic spline analysis indicated that risk of adverse events stabilized 9 months after stroke. Recurrent stroke rates are elevated even in patients not having surgery, but a subsequent analysis by the authors showed that surgery magnifies this risk.16
Timing of Emergency Surgery after Stroke: Christiansen et al. (Denmark)12
A subsequent study of the Danish registry data examined rates of 30-day MACE and mortality for emergency surgeries performed after stroke. The design and exclusion criteria mirrored the earlier study of elective surgeries, but the intervals were stroke <3 months before surgery, 3 to 9 months before surgery, and >9 months before surgery. The elevation in risk for the <3-month group was less pronounced for the emergency surgeries compared with those seen in the prior study of elective surgery, likely reflecting the increased baseline risks of emergency surgery. In contrast to the trends seen for elective surgery, risks in the 3- to 9-month and >9-month groups were quite similar. Spline analyses suggested that risks stabilized 5 months after ischemic stroke. Risks of general anesthesia compared with other modalities were similar. Surprisingly, risks were higher for anesthesia durations of <120 minutes compared with 120 minutes or more (aOR 6.69 [95% CI: 5.44-8.23] vs. 3.93 [95% CI: 3.39-4.56] in the <3-mo group). Although not proven, the authors speculated that greater preparation and vigilance were used for longer, more involved surgeries.
Christiansen and colleagues also analyzed patients who underwent emergency surgery within 14 days of an ischemic stroke. A total of 323 patients who had surgery 4 to 14 days after stroke were compared with a propensity-score matched group of 323 patients who had surgery 1 to 3 days after stroke. Mortality rates were similar in both groups, but MACE rates were lower in patients who had surgery within 3 days (69 vs. 93 events, P=0.029).
Timing of Elective Surgery After Stroke: Glance et al. (United States)10
This 2022 study examined nearly 6 million patients at least 66 years old from a Medicare database who were admitted for elective noncardiac, non-neurologic surgery. The reference group was patients without a stroke in the prior 720 days. Their primary analysis subdivided the initial 90-day period by month (≤30 d, 31 to 60 d, and 61 to 90 d). They also stratified by high (>1%), intermediate (0.5% to 1%), and low (<0.5%) surgical mortality risk. Joint arthroplasties accounted for nearly all the low-risk surgeries in their analysis. High-risk surgery was more common in patients with a history of stroke (39.3% vs. 19.8%).
The risk of recurrent stroke after surgery exhibited distinctive temporal patterns in the high- and intermediate-risk surgery cohorts. Among high-risk patients, the adjusted ORs for recurrent stroke were 7.66 (95% CI: 5.88-9.99) in the ≤30 days group, 5.27 (95% CI: 4.07-6.83) in the 31 to 60 days group, and 3.61 (95% CI: 2.61-4.98) in the 61 to 90 days group. Adjusted ORs were similar in the 91 to 180 days (3.55, 95% CI: 2.85-4.41) and 181 to 360 days (3.37, 95% CI: 2.85-3.99) groups. In contrast, in the intermediate-risk cohort, the adjusted ORs were stable across the initial 3 months (9.0 for ≤30 d [95% CI: 5.20-15.59], 8.90 for 31 to 60 d [95% CI: 5.62-14.11], and 8.64 for 61 to 90 d [95% CI: 5.64-13.24]) and showed a stepwise decline thereafter (5.66 for 91 to 180 d [95% CI: 4.19-7.64] and 4.07 for 181 to 360 d [95% CI: 3.10-5.34]). Among low-risk (primarily joint arthroplasty) patients, the adjusted ORs for recurrent stroke were much higher at 91 to 180 days (12.19, 95% CI: 9.96-14.92) than at 181 to 360 days (7.68, 95% CI: 6.53-9.05). These results suggest that, where possible, high-risk surgeries should be postponed 2 to 3 months after stroke versus 6 to 12 months for lower-risk procedures.
Overall, the risk ratios for recurrent stroke were much lower than those observed by Jørgensen and colleagues. Part of this may be attributable to lower risks in the Danish reference group, which included younger patients and ambulatory surgeries. Glance and colleagues did not report any cubic spline analysis, so direct comparison to the Danish splines is not possible.
Timing of Stroke After Surgery: Chalitsios et al. (England)13
The most recent 2024 cohort study captured over 21 million adult patients admitted for elective or emergency surgery from a comprehensive database of National Health Service (NHS) hospitals in England. Their exposure was prior cardiovascular events (hemorrhagic, ischemic, and unspecified strokes as well as TIA and cardiac events). They reported all-cause mortality over several intervals. Surgeries were classified by invasiveness (mild, moderate, major), and their analysis included endoscopies and interventional radiology procedures in the minor category. Patients with a cardiovascular event >24 months before surgery served as a reference group for the regression models, and the median interval between the event and surgery served as the reference for cubic spline analyses.
The composite cardiovascular exposure used for the regression analysis limits direct comparisons to prior studies from Denmark and the United States. However, Chalitsios and colleagues reported cubic spline analysis of 30-day mortality after ischemic stroke (>160,000 patients) stratified by invasiveness and elective versus emergency surgery. Elective surgeries showed the highest initial adjusted OR for mortality (∼12) and risks stabilized around 9 months. Adjusted ORs for emergency surgery were much lower (<3) and stabilized roughly 3 months after ischemic stroke. Initial adjusted ORs were similar for minor and major surgeries (3-5) and stabilized by 6 months after ischemic stroke. Moderate surgeries were associated with higher adjusted ORs (initially ∼7) and stabilized 7-8 months after ischemic stroke.
Cubic spline analyses for TIA (>90,000 patients) were reported in the supplementary material. Adjusted ORs for TIA were lower (≤2) than those for ischemic stroke. CIs were wide, but risk appeared to stabilize later for TIA: ∼18 months for minor surgery, ∼15 months for moderate surgery, and ∼12 months for major surgery.
Luney et al. (England)14
This follow-up study of the English dataset used the same methodology but focused on other outcomes, including acute MI and recurrent stroke within 1 year of surgery. As before, most of the analysis used a composite of cardiovascular events as the exposure, but cubic spline analyses for prior ischemic stroke and TIA were presented in the supplementary material.
Adjusted ORs for recurrent stroke were 2 to 3 in the first month after stroke and declined rapidly over the first year. Risks leveled off more quickly for major surgeries and emergencies, with the 95% CI crossing the reference value just before 12 months. Risks for elective surgeries took ∼18 months to stabilize. After TIA, the risk of stroke remained elevated for more than a year for elective surgery.
Unanswered Questions/Future Research
It is challenging to study the perioperative risks of surgery after ischemic stroke through prospective observation and adequately powered randomized trials would be prohibitively expensive. Retrospective studies of large databases provide useful information but they are inherently limited. For example, do the risk profiles observed in admitted Medicare and NHS patients apply to patients presenting for ambulatory surgery? Surgeries that are time sensitive may have different risk profiles than those that are purely elective, but distinguishing between those categories may not be possible in administrative datasets. Database studies are also ill-suited to assess risk mitigation strategies. For example, Jørgensen and colleagues found MACE rates were lower in patients with prescriptions for anticoagulants but future research is required to define optimal perioperative anticoagulation protocols in patients with prior stroke.
Retrospective studies also cannot capture information about covert stroke, which appears to be ten times more common than overt perioperative stroke. Although detectable only by imaging, these covert infarcts carry important implications for brain health.17 Whether or not delaying surgery after stroke decreases the risk of covert stroke and improves outcomes requires further prospective study.
Future research should include both cubic spline analyses and multiple regression models. Whenever possible, the time windows in the regression models should mirror prior studies to facilitate direct comparison. It is also essential that future studies account for relevant confounding variables such as age, comorbidities, and surgical risk. For example, a recent study of total knee arthroplasty18 that reported very high unadjusted ORs for recurrent stroke is difficult to compare to the adjusted arthroplasty outcomes reported by Glance and colleagues. It would also be useful for future models to incorporate stroke features such as severity and territory, though that may be challenging to extract from administrative databases.
CONCLUSIONS
Scheduling surgery after ischemic stroke should begin with joint discussion among the patient, surgeon, and anesthesiologist (Fig. 1). The first factor to consider is whether the surgery is emergent, time sensitive, or purely elective. Emergency surgeries should proceed after informing patients of the risks involved. In the immediate aftermath of a stroke, it may be preferable to operate within 3 days if the surgery cannot be postponed several months. For cases that are purely elective, the data from England suggest that risk of recurrent stroke remains elevated for at least 18 months, so patients must weigh the benefit of delay against any quality of life benefits the surgery may offer.
When surgeries are time-sensitive, procedure-specific risk and invasiveness should be incorporated into the scheduling decision. For more invasive procedures or those with high underlying risk of mortality, most studies found minimal benefit of delaying beyond 3 months after stroke. Surgeries with intermediate risk or invasiveness showed stepwise decreases in risk over time. These procedures require careful consideration of the potential to decrease perioperative complications versus the risks of delaying surgery.