Time elapsed after ischaemic stroke and risk of adverse cardiovascular events and mortality following elective non-cardiac surgery

 Journal of the American Medical Association. 2014;312(3):269-277

 Presented by: Dr Anthony Byford-Brooks


  • Cerebrovascular accident (CVA) is recognised as a major risk factor for major adverse cardiac event (MACE) following non-cardiac surgery. It is included in the Lee revised cardiac risk index, and is also analogous to the perioperative risk to patients with recent myocardial infarction (MI) +/- percutaneous coronary intervention (PCI). The ‘safe’ timing of surgery post-CVA is not well defined like it is with MI.
  • Cerebral autoregulation is known to be impaired up to 90 days following CVA, but the significance of this in the perioperative setting, and the impact of surgery and anaesthesia on autoregulation is not well studied.
  • This study aimed to look at safety and importance of time between stroke and surgery.

Design & Setting

  • A retrospective, Danish nationwide cohort study from 2005-2011.
  • All patients above 20 years of age undergoing elective non-cardiac surgery (n=481,183).
  • Danish healthcare system keeps all information on national registries, and five were accessed in order to gather data on patient backgrounds, types of surgeries, medicines used and anaesthetic records.


  • All patients over 20 years of age having elective non-cardiac surgery from 2005-11. ICD-10 used to identify those patients who suffered ischaemic stroke in the past. Haemorrhagic stroke and TIA were excluded. Stroke diagnosis excluded if time elapsed to surgery >5yrs.
  • Five population groups, made to be roughly analogous with post-MI risk:
    1. No prior stroke
    2. Stroke ❤ months of surgery
    3. Stroke 3-6 months of surgery
    4. Stroke 6-12 months of surgery
    5. Stroke >12 months of surgery
  • Grouping for use of pharmacological agents based on use of antihypertensives, antithrombotics, oral hypoglycaemics and diuretics.
  • Significant comorbidities included organ failures, AF, IHD, COPD, PVD, anaemia, DM and metastatic disease
  • Surgery performed subdivided by specialty, excluding trauma, intracranial surgery, tracheostomy/gastrostomy and urgent upper GI. Also grouped into low, intermediate and high risk surgeries.


  • Primary outcomes were all-cause mortality and MACE within 30 days.
  • MACE subdivided into nonfatal acute MI, nonfatal ischaemic stroke and cardiovascular death.


  • Of the 481,183 surgeries performed, 7137 (1.5%) were performed in patients with prior history of stroke.
  • These patients were on average 16 years older, male, on cardiovascular meds and had more comorbidities.
  • Almost a quarter of the stroke and non-stroke group had >1 surgery in the 5 year period.
  • Incidence rates (Stroke vs. Non-stroke group):
    • Ischaemic stroke: ❤ months – 149.6x higher
    • All-cause mortality: ❤ months – 12.6x higher
  • Odds Ratios:
    • 30-day MACE (<3months vs >12 months): 14.23 vs 2.47
    • Low (9.96), Intermediate (17.12) and Higher (2.97)
    • Recurrent strokes ❤ months: 67.6
    • No association between prior stroke and acute MI
    • Cardiovascular death: 4.35
    • Splines for OR levelled off roughly after 9 months.
    • Alcohol and smoking as covariates altered OR very little.
    • Use of blood-thinning agents and statins has a significant impact in reducing risk.
  • Relative Risk:
    • 30-day mortality: 1.8-fold increased risk in stroke group.
    • 30-day MACE: 4.8-fold increased risk in stroke group.
  • Stroke patients with AF at less risk than those without AF.
  • Those with recurrent strokes at higher risk.


  • Elective non-cardiac surgery <9 months after stroke carries significant risk of MACE and mortality.
  • Low or intermediate surgery carries equal or higher relative risk than high-risk surgery.
  • Patients with AF have lower risk perhaps due to nature of stroke (thrombotic vs atherosclerotic) and higher likelihood of subsequently being on appropriate drugs. 


  • Good access of the Danish databases to address a question not previously asked.
  • Large cohorts.
  • Use of data analysis that helps quantify risk over a protracted period.
  • Consideration of additional factors (drug history, comorbidities).
  • Authors identify study weaknesses.


  • No data on whether surgeries performed took time elapsed from stroke into consideration.
  • Patients may not have been fully worked up or optimised prior to surgery, particularly those <3months of stroke e.g. echo.
  • No data on in-hospital drug administration, only long-term meds at home
  • Undiagnosed comorbidities would skew data.
  • Guidelines for perioperative use of antithrombotics in Denmark changed during study period.
  • Not able to determine if ischaemic stroke is embolic or atherosclerotic e.g. AF vs PVD.
  • Type and conduct of anaesthesia not accounted for.


Although stroke is known to be a risk factor, a way of quantifying risk based on time since event, as well as what patient factors adjust this risk, can improve patient safety. The timing is similar to acute-MI +/- PCI for those at highest risk, but some risk does remain up to 9 months.

Potential for Impact

There is a potential for impact here. A forming evidence base on the degree of risk following stroke could help mitigate adverse events by allowing for optimisation of patients following stroke, or delaying of surgery, particularly those of lower risk. It also gives food for thought for emergency surgery and highlights the need for a detailed history of CVA and informed discussion on perioperative risk for those patients.