Introduction

Over the last three decades, tremendous improvement has been achieved in emergency treatment of adult stroke with faster recognition, decreasing time to treatment and new treatment options. Systems of care which include predefined prehospital and emergency department protocols and diagnostic algorithms [1] have increased access to hyperacute interventions which improve outcomes by salvaging viable brain [2]. Unfortunately, children are not receiving the benefits of these treatments due to inordinate delays related to poor recognition of stroke symptoms and limited access to emergent imaging [3, 4, 5••, 6, 7].

Although childhood stroke has gained increasing attention, little still is known on treatment in general and especially emergency treatment of children with stroke. Existing guidelines [8, 9•, 10•] rely largely on expert opinions and extrapolation of adult evidence. However, it is well established that childhood stroke differs in aetiology [11–15]; thus, treatment should be tailored to specific paediatric needs.

This review will provide an overview of the importance of clinical recognition of childhood stroke by paramedics and emergency staff, compare adult bedside stroke recognition tools and discuss their suitability in children. In a second part, the stroke chain of survival focusing on important components of the emergency assessment of childhood stroke will be discussed. Last but not least, current knowledge and evidence on acute stroke treatments in children will be reviewed.

Clinical recognition of childhood stroke

Symptoms and signs of stroke

Most symptoms and signs in childhood stroke are similar to those of adult stroke with the exception of seizures. Prospective population-based studies suggest similar frequencies of headache, hemiparesis and facial weakness [14–17] but increased frequency of speech disturbance and seizures when compared to adults (Table 1). The frequent occurrence of seizures in childhood stroke is a notable difference to adults, occurring in up to 50 % of cases [14, 18–20]. Particular challenges to diagnosis include (a) the greater difficulty identifying posterior circulation strokes because they are often associated with non-specific signs such as headaches, nausea and vomiting, (b) children having difficulty articulating their symptoms and (c) clinical assessment being sometimes affected by limited ability to cooperate with formal neurological examination (especially important for evidence of ataxia, mild facial palsy and visual problems).

Table 1 Comparison of symptoms and signs at manifestation
Full size table

Differential diagnosis of stroke in children

Stroke is the cause of focal neurological symptoms in approximately three quarters of adults. In contrast, there is a low a priori probability of stroke in children, accounting for 7 % of cases in a recent Australian study [21••]. A wide variety of conditions mimic stroke [21••, 22, 23] which differ from adults. Migraine is the most common mimic diagnosis [21••], but other conditions include seizures, posterior reversible leucoencephalopathy, intracranial infections, demyelinating disorders, metabolic stroke, tumours, drug toxicity, conversion disorder problems and syncope.

Public education campaigns

Knowledge of stroke symptoms in the community, amongst paramedics, primary care and emergency physicians, is essential to decreasing diagnostic delays. Public education campaigns such as Face-Arm-Speech-Time (FAST) and SUDDENS [the five sudden stroke warning signs: 1) weakness or numbness, 2) troubles talking or confusion, 3) troubles seeing, 4) troubles walking, imbalance or dizziness; 5) sudden headache] have been developed to improve lay awareness of stroke. The FAST message [24] relies on three symptoms: facial palsy, hemiparesis and speech problems. FAST is less sensitive than the SUDDENS message capturing additional symptoms of severe headache, focal numbness, visual disturbance, dizziness, trouble walking, loss of balance or coordination. SUDDENS is more sensitive than FAST but more difficult to remember [25]. Therefore, many countries around have adopted the simpler FAST message.

Bedside stroke recognition tools

Paramedics, triage nurses and emergency physicians are the first point for many patients with stroke [26], and bedside stroke recognition tools have been developed to assist these health professionals to identify stroke. Paramedic tools include the most widely used FAST [24] and the Cincinnati Prehospital Stroke Scale (CPSS) [27]—both shown to have the best operating characteristics in a systematic review [28]. Emergency physician tools include the Recognition of Stroke in the Emergency Room (ROSIER) tool [29]. Most tools include “absence of seizures” as a variable to increase specificity by identifying mimics. Thus, these tools may have limited utility in children because seizures are a frequent presenting symptom of stroke. A comparison of FAST and ROSIER in a paediatric emergency setting revealed both tests to have reasonable overall sensitivity (78 and 81 %, respectively) [30••]. The ROSIER was better for posterior circulation stroke, which accounts for approximately one third of all events [31, 32]. Specificity of bedside tools needs to be determined in prospective paediatric cohorts.

The Stroke Chain of Survival: improving access to hyperacute therapies

There is a major focus in adults on reducing delay to stroke diagnosis because thrombolysis is established as an effective treatment if given up to 6 h from symptom onset [2]. The Stroke Chain of Survival and Recovery is a framework devised by the National Institute of Neurological Disorders and Stroke (NINDS) to improve acute stroke care by identifying key points at which delays occur which can be targeted by interventions [33, 34]. Adult emergency stroke management guidelines recommend radiological confirmation of diagnosis within 45 min, decision to thrombolyse within 60 min and admission to a stroke unit within 3 h [33, 34].

Understanding of reasons for delay along the paediatric Stroke Chain of Survival, from symptom onset to radiological diagnosis, is essential to improve outcomes for children affected by stroke. Prehospital factors, such as failure to call an ambulance, are important contributors to delayed diagnosis in adults [35]. In contrast, in-hospital factors contribute more to delayed diagnosis in children. Limited access to diagnostic imaging is probably the biggest obstacle to rapid stroke diagnosis in children [5••].

Potential interventions in the prehospital setting include the parental education, particularly for children affected by congenital heart disorders or sickle cell disease who are at increased risk of stroke. Development of primary paediatric stroke centres with well-developed diagnostic and acute management protocols will increase access to hyperacute therapies [36].

Assessment and diagnosis of stroke

Clinical evaluation of children with suspected stroke

Similar to adults, predefined pathways adjusted to different local in-hospital facilities need to be in place. Primary survey should include assessment of vital signs, including blood pressure and saturations, and level of consciousness. Assessment of symptom severity should be incorporated into the neurological examination. The NIH stroke scale which has been recently validated in children pedNIH scale [37] provides an overall impression of the severity of the clinical situation, and can be used to monitor clinical course. Stuttering and worsening of symptoms are frequent in childhood-onset stroke and may be a clue to specific aetiologies such as focal cerebral arteriopathy [38]. In contrast, embolic strokes typically reach maximum severity immediately following symptom onset.

Neuroimaging

Once a child has arrived in emergency, immediate investigations are required to confirm the clinical suspicion of stroke and initiate treatment. CT imaging to identify intracranial haemorrhage prior to consideration of thrombolysis is the standard adult approach to stroke diagnosis. MRI is the investigation of first choice in children with suspected ischaemic stroke for several reasons (Fig. 1). The low probability of stroke in children means it has to be positively diagnosed on imaging [21••, 22, 23]. CT imaging has poor sensitivity for detection of acute ischaemia, ranging from 16 to 56 % in paediatric studies [6, 7, 39, 17, 5••]. As summarized in Fig. 1, MRI has also the potential to answer further important questions concerning vessel status [40], diffusion-perfusion mismatch (important for decision on lyses) [41], search for dissections [42•] and inflammatory signs [43]. Equally important is that MRI is more sensitive for detection of some mimic diagnoses including demyelinating disorders, encephalitis and tumours, particularly those involving the posterior fossa. Some of this information can be obtained by modern CT scanners [44], but there are concerns about the lifelong risk of malignancies with exposure to CT imaging in children [45, 46]. Accessing MRI imaging in the emergency setting is challenging, as evidenced by a recent study by Mallick et al. [5••] where median time to confirmation of stroke diagnosis was 23 h. However, recent adult studies have demonstrated the feasibility of rapid MRI protocols in the ED [47•].

Fig. 1
figure 1

Neuroimaging for children with suspected stroke. DWI diffusion weighted images, ADC apparent diffusion coefficient, CE contrast enhanced, MRA magnetic resonance angiography, SWI susceptibility weighted images, TOF time of flight, SE spin echo, IR inversion recovery. *May be done in a second examination within the first few days. **In case of considering thrombolysis.

Full size image

Laboratory investigations

Basic blood work in the emergency department should include full blood count, electrolytes, glucose and renal and liver function. Further investigations as summarized in Table 2 might be important to initiate immediately after diagnosis (as signs for infectious or inflammatory aetiology); others can/have to be done at a later stage (as prothrombotic disorders). There is increasing data to support the importance of CSF analysis in childhood stroke [48], although it is contraindicated in children requiring anticoagulation treatment. Indicators of infectious or inflammatory processes include elevated opening pressure, pleocytosis and elevated protein [49].

Table 2 Suggested laboratory investigations
Full size table

Acute treatment of stroke

Treatment recommendations in published paediatric stroke guidelines are largely based on expert consensus opinion or extrapolation from adult data due to a lack of evidence from randomized controlled trials [8, 10•].

General principles of stroke care

In any child with stroke, it is important to secure the basic ABC of airway, breathing and circulation. In contrast to children with haemorrhagic stroke, a significant reduction in level of consciousness is unusual in children with ischaemic stroke [14, 17]. There is conflicting evidence to support oxygen therapy by mask breathing in adults with normal saturations (≥92 %) [50, 51]. No such treatment seems to be justified in children, in the absence of significant cardiac disease, but careful monitoring and maintaining of oxygen saturations ≥92 % is important and recommended in adult [1] and paediatric [8] guidelines. Arterial hypertension is a major risk factor for stroke in adults, and many adults have elevated blood pressure at presentation [52]. Elevated blood pressure and rapid reduction of blood pressure both have negative effects on outcome. In view of this equipoise, adult stroke guidelines propose an optimal blood pressure to be in the mild to moderately elevated range, based on data from randomized controlled trials [1, 53]. Elevated blood pressure is reported in approximately two thirds of children in the first 24 h following stroke onset, and in about 20 % during the first 3 days, mainly in children with cardiac problems, Moyamoya and severe occlusive vasculopathy [54, 55]. It is possible that the elevated blood pressure may be a compensatory phenomenon to maintain cerebral perfusion, but increasing hypertension may also be the first sign of raised intracranial pressure. Elevated blood pressure in the first few days should result in assessment for malignant cerebral oedema, haemorrhagic conversion and cardiac or renovascular disease. Correction of blood pressure should be done cautiously in line with adult recommendations. Maintenance of euvolemia in patients with stroke is important, for adults as well as for children. Hypovolemia might result in hypoperfusion and thrombocytosis, and promote thrombosis. In children at risk of malignant swelling of infarction, careful monitoring of fluid balance is necessary and, for intravenous fluid, isotonic solution is important. Pyrexia is associated with increased short-term mortality in adults [56]. Paracetamol is recommended for adults with fever [57]. A similar approach is recommended for children even though no specific paediatric data are available [8]. For pre-schoolers, treatment for fever may also reduce the risk of acute symptomatic seizures. It makes common sense to treat infectious foci by specific antiinfectious agents. For adults, hypo- and hyperglycaemia should be avoided [58, 59], because persistent hyperglycemia leads to increase of infarct volume and cerebral oedema, both negatively affecting outcome. Significant hypoglycaemia is known to provoke seizures. Therefore, blood sugar monitoring is recommended in the acute phase following stroke onset. Without accompanying diabetic or metabolic disorders, glucose problems in children are rare.

Acute treatment options

Despite lack of any evidence, there is in large a basic agreement on how to treat children with stroke (Table 3) [9•, 10•]. However, individual decisions have to be made carefully, considering not only known evidence in adults but also the special aspects in each child.

Table 3 Medical treatment options
Full size table

Thrombolysis and mechanical thrombectomy

The ischaemic penumbra, which surrounds the infarct core, is the target for hyperacute interventions. This hypoperfused tissue, which is functionally inactive but structurally intact, can be salvaged with recanalization. The time characteristics of the penumbra have not been adequately studied in children. However collateralization, which has an important influence on infarct growth, is thought to be superior in children [60–62, 41] and may mean there is a longer time window for intervention.

In adults, intravenous thrombolysis and intra-arterial lyses improve outcome if initiated within time limits of 3.5–4 and 6 h, respectively [63–65]. Time to recanalization is one of the important predictors of outcome [41]. Recanalization is more often achieved by intra-arterial thrombolysis than intravenous thrombolysis [66]. For these reasons, bridging intravenous thrombolysis, followed by angiography and additional intra-arterial thrombolysis, is becoming the treatment of choice in adults [67]. There are increasing data to support expanding the treatment time window and offering treatment to elderly patients [68]. The recently published study MR CLEAN has shown that mechanical thrombectomy is superior to lyses by tpA or urokinase [69].

There are no controlled data to support thrombolysis or thrombectomy in children. But there are increasing reports of off-label usage of tPA in children. In a review of the literature summarizing 17 cases after thrombolysis (iv or ia), outcome was successful in 71 % [70]. In another study of 9257 children, only 0.7 % received thrombolysis [71]. A review of centres contributing data to the International Stroke Registry [72] found that 2 % of 687 children received tPA but treatment often did not comply with recommended adult guidelines. Time to administration ranged from 2 to 52 h for intravenous and 3.8 to 24 h for intra-arterial tPA. Symptomatic intracranial haemorrhage occurred in 26 % of patients, which is much higher than the 7 % bleeding rates reported in adult trials [2]. Outcome of these children was not favourable, with two children dying and 12/15 survivors having residual neurological deficits [72]. When results were compared to ten cases previously described in the literature, children in the registry were younger (p = 0.07), treatment was more often delayed (p = 0.057) and outcome was poorer, suggesting a previous publication bias towards cases with shorter treatment lag and better outcomes.

The most appropriate dose of tPA in children is unknown. Developmental differences in the fibrinolytic system such as lower plasminogen and free tissue plasminogen levels and higher plasminogen activator inhibitor-1 (PAI-1) levels may mean that increased tPA doses are required relative to adults to promote clot lysis [73, 74]. There may also be age-related differences in risk of bleeding complications. For example, a recent thrombolysis study in young adults aged 16 to 49 years found lower complication rates when compared to major RCTs. In particular, none of the 48 young adults treated developed symptomatic intracranial haemorrhage [75•].

Emerging evidence in adults suggest mechanical thrombectomy to be the most effective means of achieving recanalization in adults [69]. A recent literature review [76••] summarized endovascular approach in 34 children. Despite delayed treatment (mean 14 h, range 2–18 h) partial or complete recanalization was achieved in 63 %. Periprocedural complications were present in 29 %, but symptomatic haemorrhages occurred in only one child. The role of mechanical thrombectomy remains unclear, and specific challenges in children include small vessel size and frequent inflammatory reaction within the vessel wall.

In summary, there are no controlled data to support thrombolysis or mechanical thrombectomy in children with ischemic stroke. However, there is probably a subgroup of children who may benefit but treatment should be considered very cautiously and only be performed in primary stroke centres experienced in acute management of children affected by stroke [36]. In the review by Ellis et al. [76••], there are cautious suggestions on how to proceed for thrombolysis/thrombectomy in children.

Aspirin versus heparin: an ongoing debate

In children where thrombolysis is not indicated, antithrombotic treatment should be initiated immediately following confirmation of diagnosis. In adults, evidence from large randomized controlled trials supports use of aspirin over heparinoids to improve outcome and reduce early recurrence without increasing risk of early haemorrhagic complications [77]. Recent AHA guidelines suggest aspirin as first-line treatment, if lyses or endovascular approach is not indicated [1]. For children, no such trials exist. A theoretical debate on anti-inflammatory and antithrombotic effects of acetylsalicylic acid versus heparinoids would suggest to favour the second. An international analysis of current practice found that aspirin, anticoagulation and no treatment were being used equally, with a trend to platelet aggregation in Europe and to anticoagulation in North America [78]. Low molecular weight heparin and clopidogrel are being increasingly used in children and are shown to be safe [79–82]. In the absence of data from randomized controlled trials, there is non-consensus amongst experts, with one treatment guideline suggesting it is reasonable to treat initially with heparinoids until exclusion of cardioembolic cause and extracranial dissection [9•], whereas the other recommends either heparinoids or aspirin as initial therapy [10•]. For adults, it has been shown that anticoagulation is superior to aspirin for reduction of recurrent cardioembolic stroke [83]. However, a single randomized controlled trial of children with complex cyanotic heart disease did not show that heparin was superior to aspirin for children for prevention of thrombotic events following palliative surgery [84]. However, both paediatric guidelines favour use of anticoagulant treatment in children if stroke is due to cardioembolism or dissection.

No randomized controlled trials exist for dissection in adults or children, but current guidelines suggest it reasonable to anticoagulate children with stroke due to extracranial dissection [1, 42•, 8, 9•, 10•]. For adults and children, there are increasing reports of endovascular approach (stenting or ballooning) in dissection problems refractory to treatment [85–87, 42•].

Decompressive craniotomy

Large middle cerebral artery infarctions and cerebellar infarctions are at risk of swelling during the first 72 h, increased intracranial pressure typically manifests by increased blood pressure, followed by secondary herniation and brainstem compression causing deteriorating consciousness and risk of death. In two paediatric series, 12.5 % of children with middle cerebral artery strokes [88] and 11 % of posterior circulation strokes [89] developed malignant cerebral oedema. Malignant middle cerebral infarction and cerebellar infarction are associated with poor outcome, and small series suggest decompressive surgery may be of benefit [90•, 91], similar to what was described in adults [92, 93]. Thus, careful monitoring of patients at risk of malignant infarction in a paediatric intensive care unit with access to emergency neurosurgical treatments is mandatory.

Steroids and immunosuppression in inflammatory problems

Infection and inflammation play an important role in stroke pathogenesis, particularly in children with cerebral arteriopathy [94, 95, 80]. Recent studies [96, 97] have identified changes in inflammatory biomarker profiles in children in the acute phase following stroke, some of which may be indicators for increased risk of recurrent event. Modern neuroimaging techniques showing vessel enhancement may provide supporting evidence of inflammation in some children with focal arteriopathies [43]. Reported recurrence risk in focal arteriopathy varies from 15 to 45 %, possibly with the highest risk in post varicella angiopathy [38, 95, 94]. For these reasons, adjunctive immunosuppressive agents are sometimes considered as “disease-modifying” therapy in addition to antithrombotics. Steroids might alter the natural history of focal arteriopathies and decrease the risk of recurrence. Benseler et al. found there was decreased recurrence risk in children with non-progressive large to medium vasculitis, who were treated with corticosteroids and antiplatelet therapy, when compared to those treated with antiplatelet treatment alone (11 vs 30 %) [49]. Further studies are required to demonstrate a beneficial effect of corticosteroid in children with focal self-limited vasculopathy. As varicella is frequently implicated as a cause of focal cerebral arteriopathy, additional treatment with acyclovir should also be considered, especially in children with positive VZV PCR in CSF. High-dose corticosteroids, followed by other immunosuppressive agents including mycophenolate, azathioprine or cyclophosphamide, are also used in children with suspected primary CNS vasculitis, which is typically associated with small-vessel disease.

Future directions

It is vital to develop predefined institutional acute paediatric stroke management protocols [36] because they have been shown to reduce the time to neurological assessment and increase thrombolysis rates in adults [98, 99]. These protocols should include activation of a “Code Stroke” with rapid evaluation by a paediatric stroke neurology team. Support from experienced adult stroke teams is also of value. Development of thrombolysis guidelines, which conform to standard adult inclusion and exclusion criteria, is essential to select appropriate candidates and minimize risk of complications. For example, adopting standard adult criteria of a paediatric National Institutes of Health Score of ≥4 is reasonable because a recent study comparing children to young adults showed similar initial stroke severity, which correlated with the severity of later deficits, thereby disproving the longstanding belief that childhood stroke is milder and has better outcome than that of adults [100•].

There is a substantial body of evidence demonstrating the positive benefits of stroke units in reducing length of hospital stay and institutional care [101] and stroke-associated morbidity and mortality when compared to care on conventional medical wards [102, 103]. It is not practical to develop geographically distinct stroke units within paediatric hospitals because stroke is uncommon, but there is no reason why the general principles of stroke unit care cannot be applied to children. The main obstacles to development of such services are the identification and support of dedicated medical leads, development of stroke emergency algorithms, funding of stroke care coordinator positions and development of standardized paediatric stroke management guidelines.

The catch cry “Time is brain” is equally applicable to children. The social and economic burden of disease for children affected by stroke and their families is substantial when considered over a life which may span 50–70 years. The best way to improve outcome is to minimize extent of injury by salvaging at risk brain. Therefore, there needs to be increasing focus on developing strategies to improve access to treatments which have transformed the approach to stroke care in adults.