Postoperative stroke

Editor, The mortality rate for postoperative stroke exceeds 25%1 even though managed by multidisciplinary teams. The overall incidence of postoperative stroke (0.05 to 1%) and the recent years have seen improvements in outcome.1–3 Our centre is a tertiary hospital and reference centre for stroke management. When needed, the ‘stroke alert’ is activated by which all the specialists are notified. We retrospectively analysed the computerised hospital's clinical records for 2016 to identify patients diagnosed with postoperative stroke within 30 days after surgery.2 Approval for this study was provided by the Ethical Committee of Hospital Clínic de Barcelona, Spain (President Dr Begoña Gómez) on 14 December 2017 (HCB/2017/0954). Data recorded included demographic characteristics, concomitant diseases, medications, type of surgical procedure, time of appearance of stroke during the peri-operative period, time of any activation of the in-hospital stroke alert, treatment and outcome. Postoperative stroke was diagnosed in nine patients out of a total of 22 666 surgeries. Thus, the overall incidence was 0.04%. All patients (six men and three women, age range 55 to 92 years) shared cardiovascular risk factors and were receiving either antiplatelet or anticoagulant treatment prior to surgery. Four patients had atrial fibrillation, and three had previously experienced an episode of ischaemic stroke. Surgery was scheduled in five cases. One of these patients, who had a previous history of ischaemic stroke, underwent a resection of a cerebral meningioma and antiplatelet treatment had been withdrawn 7 days prior to surgery. The four other scheduled operations were cardiovascular (three aortic valve replacements, one of which involved a coronary bypass and one mitral and tricuspid valve replacement). These patients were on anticoagulant treatment before surgery and received bridging therapy with low-molecular-weight heparin. Urgent operations accounted for the four remaining cases: a humerus fracture repair, in which apixaban had been suspended 2 days prior to surgery; a colonoscopy to study rectal bleeding, in which acenocoumarol had been suspended because of bleeding risk; management of a splenic abscess; and fibrinolysis for acute arterial ischaemia. Stroke developed within 24 h of surgery in two cases and within the first 5 days in six cases. Stroke occurred later (10 days after surgery) in a single case. The in-hospital stroke alert was activated in five of the nine cases (55.5%), but activation occurred within the time frame for treatment in only two of those cases. One of these patients received thrombolytic therapy. One patient died (11%) due to severe cerebral oedema that lead to brain death in spite of decompressive craniectomy. Five patients with poor outcome required nursing home care and three patients with better outcome were home discharged. Our review of cases, like most studies of postoperative stroke in the literature, was retrospective and conclusions must therefore be cautious. A recently published big database including more than 150 000 patients found a stroke incidence of 0.6%, 10 times higher than ours. This difference might be explained by multiple factors that are likely to be promoted by the retrospective nature of such approaches, including definition of stroke, adequate documentation and other confounding factors.3 Although most events occur within 10 postoperative days, they can also happen after discharge. As a result, patients may be attended, diagnosed and treated in other centres, in which their cases are not classified as peri-operative stroke. We did not seek the records of patients after they left our hospital. In addition, patients may experience covert stroke, in which there is no evident neurological dysfunction, only cognitive alterations that can only be diagnosed with specific tests that are not usually performed in clinical settings outside the context of research.1,4 Finally, in most conventional wards, formal neurological evaluations are not the rule, whether due to lack of staff experience or awareness of a problem or the complexity of standardised tests.1 All these data show that peri-operative stroke might be underdiagnosed in most hospitals and strengthens the conclusion that efforts for early recognition have to be implemented. Our review leads us to conclude that despite advances in stroke diagnosis and treatment, surgical patients who develop postoperative stroke are diagnosed too late and therefore do not receive proper treatment to re-establish brain circulation.6 Various causes for delay have been identified. The patients’ condition may be medically complex because of advanced age and comorbidities. Postoperative complications such as arrhythmias, bleeding or infection may be present; or other common confounding factors (sedatives, intubation, regional anaesthesia) may make early diagnosis a greater challenge after surgery than when stroke happens in the general population.5 These problems, along with the prothrombotic state associated with surgery, worsens the prognosis, bringing mortality rates as high as 25%, double than in other settings.1 Treatment options have been limited in the past as systemic thrombolysis is contra-indicated in the first 14 days after surgery (US American guidelines) or even during the first 3 months (European guidelines).6 Treatments that encourage brain reperfusion (intra-arterial thrombolysis and mechanical thrombectomy) are now available. These options, especially mechanical thrombectomy, seem the most reasonable and safe way to achieve reperfusion, but the rate of peri-operative strokes with vascular occlusion potentially accessible for mechanical thrombectomy is lacking. This reason, and because of the comorbidities and underlying diseases requiring surgeries, makes patients’ outcome worse than in strokes not associated with peri-operative period.7 In the light of the limitations of retrospective studies, it seems appropriate to demand a prospective trial or to set up a registry which would allow adequate documentation of this rare but devastating event. With our review, we conclude that new strategies are required. First, we must endeavour to identify stroke as it develops, following all patients with pre-operative cardiovascular risk factors to establish adequate prevention measures, as well as establishing criteria of severity through universal scales such as National Institute of Health Stroke Scale, NIHSS scale. An in-hospital multidisciplinary system for early recognition and diagnosis of this problem should be able to raise awareness and provide training for staff members who take charge of these patients. Faster transfer of patients to the specialised team, with more treatment possibilities that can improve prognosis, is key. Acknowledgements relating to this article Assistance with the letter: none. Financial support and sponsorship: none. Conflicts of interest: none.