Severe acute ischemic stroke (AIS) has been recognized as deserving management in the neurocritical care unit (NCCU) and considerable progress in its understanding and management has been made over the last 10 years. The results of older studies had put into question the usefulness of giving patients with severe AIS access to NCCU treatment and mechanical ventilation, based on a very poor reported prognosis with mortality rates between 50 and 80% [1–4]. Today, however, treatment options such as endovascular thrombectomy, decompressive surgery, or targeted temperature management have changed the perspective of these patients, and they require rapid, adequately aggressive, and consequent emergency and critical care. This chapter addresses, in fairly chronological order, the step-wise management of severe ischemic stroke, i.e. the acute assessment, stabilization, and recanalizing treatment in the emergency room (ER), the general aspects of care for ischemic stroke patients in the NCCU, and finally specific treatment aspects associated with different types of ischemic stroke. Other features of stroke care, such as recognition and prehospital management, general stroke unit management, and secondary stroke prophylaxis, will not be covered.
Severe acute ischemic stroke (AIS) has been recognized as deserving management in the neurocritical care unit (NCCU) and considerable progress in its understanding and management has been made over the last 10 years. The results of older studies had put into question the usefulness of giving patients with severe AIS access to NCCU treatment and mechanical ventilation, based on a very poor reported prognosis with mortality rates between 50 and 80% [Reference Berrouschot, Sterker, Bettin, Koster and Schneider1–Reference el-Ad, Bornstein, Fuchs and Korczyn4]. Today, however, treatment options such as endovascular thrombectomy, decompressive surgery, or targeted temperature management have changed the perspective of these patients, and they require rapid, adequately aggressive, and consequent emergency and critical care. This chapter addresses, in fairly chronological order, the step-wise management of severe ischemic stroke, i.e. the acute assessment, stabilization, and recanalizing treatment in the emergency room (ER), the general aspects of care for ischemic stroke patients in the NCCU, and finally specific treatment aspects associated with different types of ischemic stroke. Other features of stroke care, such as recognition and prehospital management, general stroke unit management, and secondary stroke prophylaxis, will not be covered.
What is a “Severe” Ischemic Stroke?
Most forms of ischemic stroke can and should be managed on stroke units. However, if the main stem of a brain-supplying vessel such as the distal internal carotid artery (ICA), the proximal middle cerebral artery (MCA) or the basilar artery (BA) is occluded, the affected brain territory is large and the resulting deficit substantial. Vertebrobasilar stroke can also result in impairment of vital functions such as circulation, breathing, and airway protection. Finally, ischemic stroke can have cerebral (e.g. edema) or systemic (e.g. cardiac arrhythmias) sequelae that cannot be sufficiently handled on stroke units. There is no official definition of “severe” stroke, but the following criteria might characterize this dimension (one or more): substantially disabling deficit (National Institute of Health Stroke Scale (NIHSS) >15), immediate near-complete or complete dependence (modified Rankin Scale (mRS) 4–5), compromise of vital functions, association with extra-cerebral complications, features of space-occupying effect or brainstem involvement on computed tomography (CT).
When Should a Patient with AIS be Placed in the NCCU?
Patients with ischemic stroke that display the above-named features may at first be managed in the ER or on the stroke unit, or be admitted to the NCCU for prophylactic close monitoring. While it is customary in some centers to admit all patients to the NCCU who received thrombolysis, many centers manage these patients very successfully on the stroke unit. The key point is that AIS patients must never be without a monitor in the acute phase. If AIS patients develop the following features, however, they should certainly be transferred to the NCCU (Table 11.1).
Large infarction with signs of swelling and/or mass effect on cerebral imaging
Postendovascular treatment if this involved intubation and mechanical ventilation
Need for neurosurgical operations (e.g. decompression) or invasive interventions
Instability during thrombolysis or thrombolysis-related bleeding
Compromise in airway protection with risk of aspiration
Respiratory failure and need of intubation
Substantial hemodynamic instability
Progressive decline of level of consciousness
Types of ischemic stroke often requiring NCCU care are given in Table 11.2.
Any stroke treated endovascularly requiring postinterventional intensive care
Large hemispheric stroke (“malignant” middle cerebral artery infarction)
Space-occupying cerebellar stroke
Stroke from large-vessel arterial dissection
Stroke from bacterial endocarditis
Concerning the early management of acute ischemic stroke, new guidelines from the American Heart Association/American Stroke Association have been very recently published [Reference Jauch, Saver and Adams5]. Reference to these guidelines will be made throughout the chapter. Stroke has to be recognized as an absolute emergency and treated with at least the same priority as a myocardial infarction. Although certain aspects of the ischemic stroke patient must under no circumstances be neglected during ER assessment, this has to be achieved rapidly. “TIME IS BRAIN” is not only a very true pathophysiological rationale but has been demonstrated in early stroke trials [Reference Marler, Tilley and Lu6,Reference Hacke, Donnan and Fieschi7] and confirmed in countless more recent studies. An organized approach (Table 11.3, Figure 11.1), ideally involving a protocol, code, and dedicated team, has proved beneficial and is recommended in current guidelines [Reference Jauch, Saver and Adams5].
|Detection||Prehospital recognition of stroke signs and symptoms|
|Dispatch||Prehospital immediate 911-activation emergency medical system (EMS) dispatch|
|Delivery||Prehospital immediate transport to and notification of stroke hospital|
|Door||Immediate ER triage to high-acuity area|
|Data||Rapid ER evaluation, monitoring, laboratory tests, imaging|
|Decision||Diagnosis, choice of acute treatment, discussion with patient/family|
|Drug||Application of drugs/interventions|
|Disposition||Timely transfer to stroke unit or NCCU|
Figure 11.1 Suggested approach to acute and critical care of ischemic stroke
Time frames have been suggested by panels of stroke experts (National Symposium on Rapid Identification and Treatment of Acute Stroke) (Table 11.4) to achieve thrombolysis in an eligible AIS patient within 60 min of entering the hospital.
|Door to physician||<10|
|Door to stroke team||<15|
|Door to CT initiation||<25|
|Door to CT interpretation||<45|
|Door to drug (“needle”)||<60|
|Door to stroke unit (SU)/NCCU admission||<180|
Heart rate, blood pressure, and systemic oxygen saturation have to assessed immediately in all patients suspected for AIS upon arrival in the ER. Should any of these vital functions be unstable, they need to be stabilized first. Although rarely necessary, this might even include intubation and ventilation of the patient before further assessment can be continued. A quick general examination of systems should follow, as well as assessment of the Glasgow Coma Scale (GCS).
Taking a brief history from the patient has to be attempted, particularly with regard to the main present complaint, the dynamics of its development and the point in time of the insult. The latter is the most important aspect in AIS history. If the exact time of onset cannot be clarified, it has to be estimated and the last point in time that the patient was without symptoms (“last seen well”) be taken for the time of onset to define the time-window. Often, however, patients with severe AIS are not able to provide a history themselves, either due to aphasia/anarthria or decline in level of consciousness and the key elements of the history have to be obtained from the emergency personnel or the family, if present. In addition to the aspects named above, further information should be obtained on past medical diseases and comorbidities (especially recent bleeding, trauma, or operations), current medication (especially anticoagulants), and the prehospital condition of the patient. All of these may influence the decision on acute treatment. Usually, the typical constellation in the history will be that of a sudden neurological deficit with a syndrome correlating to a brain vessel territory, with the symptoms remaining unchanged or gradually resolving. However, some types of ischemic stroke may present with stuttering or step-wise worsening symptoms, such as in basilar thrombosis, sinus/venous thrombosis, or arterial dissection.
A five-minute neurological examination of the suspected stroke patient has to include testing of the level of consciousness, speech, response to simple commands, cranial nerves II/III/IV/VI/VII, the main tendon reflexes, Babinski´s sign, limb power, and localizing response to (painful) stimuli. At times, orienting testing of coordination may by feasible, too. ER personnel may use a well-validated tool for initial examination or triage of the patient suspected for AIS, the Recognition of Stroke in the Emergency Room (ROSIER) scale. This tool comprises seven items: loss of consciousness/syncope, seizures, new asymmetric facial weakness, asymmetric arm weakness, asymmetric leg weakness, speech disturbance, and visual field defect. In the validating study, it had a sensitivity of 92%, a specificity of 86%, a positive predictive value (PPV) of 88%, and negative predictive value of 91% [Reference Nor, Davis and Sen8].
From the history and the neurological examination, one should try to rule out so called “stroke mimics,” such as seizures, migraine attacks, or hypoglycemia. However, if there remains suspicion that the patient may suffer from AIS, consideration of these mimics should not lead to withholding stroke treatment.
Grading the Deficit
The most common way to score the severity of ischemic stroke is by the National Institutes of Health Stroke Scale (NIHSS). This score has not only been the best validated and most widespread score in all large relevant stroke trials, it has also proved very helpful in monitoring the clinical course and informing treatment decisions in everyday stroke care, and its use is recommended in current guidelines [Reference Jauch, Saver and Adams5] (Table 11.5).
|1A. Level of consciousness||0 – Alert|
|1 – Drowsy|
|2 – Obtunded|
|3 – Coma/unresponsive|
|1B. Orientation questions (2)||0 – Answers both correctly|
|1 – Answers one correctly|
|2 – Answers neither correctly|
|1C. Response to commands (2)||0 – Performs both tasks correctly|
|1 – Performs one task correctly|
|2 – Performs neither|
|2. Gaze||0 – Normal horizontal movements|
|1 – Partial gaze palsy|
|2 – Complete gaze palsy|
|3. Visual fields||0 – No visual field defect|
|1 – Partial gaze palsy|
|2 – Complete gaze palsy|
|4. Facial movements||0 – Normal|
|1 – Minor facial weakness|
|2 – Partial facial weakness|
|3 – Complete unilateral palsy|
|5. Motor function (arm)|
|0 – No drift|
|1 – Drift before 5 seconds|
|2 – Falls before 5 seconds|
|3 – No effort against gravity|
|4 – No movement|
|6. Motor function (leg)|
|0 – No drift|
|1 – Drift before 5 seconds|
|2 – Falls before 5 seconds|
|3 – No effort against gravity|
|4 – No movement|
|7. Limb ataxia||0 – No ataxia|
|1 – Ataxia in one limb|
|2 – Ataxia in two limbs|
|8. Sensory||0 – No sensory loss|
|1 – Mild sensory loss|
|2 – Severe sensory loss|
|9. Language||0 – Normal|
|1 – Mild aphasia|
|2 – Severe aphasia|
|3 – Mute or global aphasia|
|10. Articulation||0 – Normal|
|1 – Mild dysarthria|
|2 – Severe dysarthria|
|11. Extinction or inattention||0 – Absent|
|2 – Mild (loss, one sensory modality lost)|
|3 – Severe (loss, two modalities lost)|
Certain NIHSS “cut-off values” have been used as a basis for treatment decisions. While it had been suggested in the past to consider an NIHSS of at least 4, reflecting a stroke relevant enough to deserve thrombolysis and to consider an NIHSS >25 a devastating stroke in which thrombolysis should be withheld, these traditions have to be regarded with great caution today. Minor strokes with an NIHSS <4 can still be disabling for some patients, may well progress later in the course, and can be quite safely thrombolyzed. Furthermore, stroke patients with an NIHSS >25 may be identified by MRI imaging as being severely hypoperfused, but not yet infarcted in a large brain territory, or may have a small infarction area in a strategically important area. The NIHSS appears very helpful, however, in identifying patients with a large vessel occlusion (LVO). In a very interesting study, Fischer et al. correlated the NIHSS of 226 consecutive AIS patients with arteriographic findings and found that an NIHSS of 12 or more predicted a central occlusion (BA, ICA, M1, M2) with a PPV of 91% [Reference Fischer, Arnold and Nedeltchev9]. This might influence decisions on advanced imaging and/or recanalizing treatment methods.
Blood sampling on admission should always include blood glucose, electrolytes, renal function studies, complete blood count with platelet count, and the basic coagulation parameters prothrombin time (PT), activated partial thromboplastin time (aPTT), and the international normalized ratio (INR). Other parameters such as troponin T, thyroid stimulating hormone, liver function tests, pregnancy test, toxicology screen, or D-dimer test may be desirable in some patients, but should not delay overall assessment. Troponin T may be elevated in AIS and often does not indicate myocardial infarction. It should therefore be interpreted in the context of the electrocardiogram (ECG) and kidney function.
Assessment of coagulation is most important in this setting, as it directly affects treatment decisions on recanalization. In patients on heparin, the aPTT, and on warfarin, the INR, respectively, will demonstrate therapeutic activity. Point-of-care devices, especially for coagulation testing (possibly the most important laboratory tests in acute ischemic stroke), can save valuable time [Reference Rizos, Herweh and Jenetzky10]. For patients on new oral anticoagulants (NOACs), reliable tests of drug activity may not be available in the ER. The ecarin clotting time (ECT) has a robust and linear relationship with levels of direct thrombin inhibitors, such as dabigatran. The thrombin time (TT) is also a sensitive indicator of the presence of these drugs, but may be influenced by other anticoagulants. A normal TT excludes the presence of significant dabigatran acitivity, while the PT/INR is not helpful. Assessment of the activity of direct factor Xa inhibitors, such as rivaroxaban or apixaban, afford specific factor Xa assays [Reference Jauch, Saver and Adams5]. It has to be stressed that this is an area of considerable uncertainty, and if the history suggests that the patient has taken NOACs within a time frame compatible with persisting therapeutic action and no tests to determine the latter are rapidly available, intravenous thrombolyis has to be refrained from and/or other ways of recanalization be considered.
Timely cerebral imaging is paramount in AIS. Recommendations for the ideal choice and sequence of imaging methods are currently under development [Reference Wintermark, Albers and Alexandrov11]. Although a large variety of imaging techniques are theoretically available, it is important for a given center to focus on and improve the locally established methods and keep in mind both practicability and time-pressure when handling AIS.
Noncontrast-enhanced computed tomography (CT) is not only widely available, quickly performed, and allows access to an unstable patient, but is excellent for ruling out intracranial hemorrhage, the main contraindication to intravenous thrombolysis (IVT). CT may also demonstrate other causes of neurologic deficits or seizures, such as large enough brain tumors, brain edema, hydrocephalus, or indirect signs of sinus and venous thrombosis. Hypodensity indicating AIS may be seen within 3–6 hours of onset [Reference von Kummer, Bourquain and Bastianello12], but earlier, small, or posterior circulation strokes may not be displayed. So-called early signs of AIS, such as loss of gray-white-matter differentiation, loss of basal ganglia distinction, cortical and insular blending, swelling of gyri with sulcal effacement, may be detected by up to 70% of observers [Reference Moulin, Cattin and Crepin-Leblond13]. They are not longer regarded as contraindications to thrombolysis. Structured scoring systems for hypodensities, such as the Alberta Stroke Program Early CT Score (ASPECTS) may improve detection of AIS [Reference Barber, Demchuk, Zhang and Buchan14,Reference Demchuk and Coutts15]. The hyperdense middle cerebral artery (MCA) sign, reflecting the thrombus, indicates LVO. The latter is a stronger predictor of severe stroke (PPV, 91%) than >50% hypodensity of the MCA territory (PPV, 75%). If frank hypodensity in more than one-third of the MCA territory is seen, thrombolysis should be withheld, since the risk of hemorrhage is high and the chance of benefits are low. Adding contrast medium for CT angiography (CTA) will display vessel occlusions and stenoses within 3–5 minutes in a quality close to conventional angiography [Reference Bash, Villablanca and Jahan16,Reference Nguyen-Huynh, Wintermark and English17], and may well influence treatment decisions. CTA is less often associated with nephropathy (2–3%) than feared in the past [Reference Krol, Dzialowski and Roy18] and offers further information on stroke age, extent, and etiology by use of source images, in combination with ASPECTS [Reference Coutts, Lev and Eliasziw19], or by thrombus evaluation [Reference Riedel, Zimmermann and Jensen-Kondering20], if these methods are established in the given center. Perfusion CT is another option to enhance information on potentially salvageable tissue and is easy to quantify. It has the disadvantages of more ionizing radiation and reduced brain coverage if scanners with a 4-cm-thick slab per contrast bolus are used. Last generation 256/320-slice scanners allow for rapid whole-brain coverage in perfusion imaging, but availability is limited.
Magnetic resonance imaging (MRI), including diffusion-weighted imaging (DWI), perfusion-weighted imaging (PWI), fluid-attenuated inverse recovery (FLAIR) sequences, and MR angiography (MRA), may be the most sensitive way to confirm ischemic stroke as early as possible, show the extent of fresh and older infarction parts, help to clarify etiology, display differential diagnoses, and, above all, display potentially salvageable tissue, all without exposing the patient to ionized radiation (Figure 11.2). It is certainly the best method of imaging for suspected brainstem stroke, in which CT is very insensitive. MRI incorporating T2*-weighted sequences is as accurate as CT in detecting cerebral hemorrhage (Reference Kidwell, Chalela and Saver21). Noncontrast-enhanced MR time-of-flight (TOF) angiography demonstrates vessel stenosis with a sensitivity of 60–85% and vessel occlusion with a sensitivity of 80–90% when compared to CTA and digitial subtraction angiography (DSA), and is thus reasonably useful in severe stroke, without exposing the patient to contrast medium (Reference Qureshi, Isa and Cinnamon22). Disadvantages of MRI are limited availability, relatively long duration of image acquisition (15–20 min), vulnerability to patient motion artifact, patient contraindications (pacemaker, claustrophobia, agitation), limited access to a potentially unstable patient during imaging, and the (low) risk of potentially dangerous nephrogenic reactions to the contrast medium gadolinium(Reference Kribben, Witzke and Hillen23) if used to enhance MRA.
Figure 11.2 Stroke MRI in MCA occlusion on the right 6 h from onset. (A) MRA displaying M1 occlusion; (B) DWI displaying small infarcted area; (C) PWI displaying larger underperfused area (= mismatch).
MRI can be recommended if the time since insult is unknown, such as in wake-up stroke. Furthermore, it is often the only way to confirm brainstem stroke and may influence treatment decisions in basilar occlusion (BAO). Most interest in MRI for AIS comes from the “mismatch” concept, i.e. the overlay of DWI and PWI sequences to delineate the penumbra, the tissue around the infarct core that is underperfused and dysfunctional, but not yet irreversibly destroyed. Many stroke centers base individiual recanalizing efforts on that concept beyond the time window. The mismatch concept for selection of well-collateralized patients for IVT within an extended time window was investigated in several studies and trials [Reference Jauch, Saver and Adams5,Reference Thomalla, Schwark and Sobesky24], which raised hopes for this approach. However, the results of two pivotal clinical trials testing the mismatch concept for IVT within the 3–9-hour window (DIAS 1 and 2) showed favorable trends, but no significant clinical benefits (Reference Hacke, Furlan and Al-Rawi25). Still, many stroke experts believe in the mismatch concept, which is considered an option for selected patients in the current guidelines [Reference Jauch, Saver and Adams5], and new studies on this concept are currently being performed, indicating possible design problems in the previous trials. Finally, MRI has been proposed for prognostication of severe stroke such as malignant hemispheric stroke with studies yielding robust results (Reference Thomalla, Hartmann and Juettler26).
Cerebral DSA is still considered the “gold-standard” of vascular imaging, but is an invasive procedure with potential risks, such as groin vessel aneurysm, bleeding, cerebral vasospasm, stroke, or even death, and considering the excellent quality of noninvasive methods such as MRA and CTA, it no longer plays a major role in early AIS assessment, unless it is used not only for diagnosis but also for subsequent endovascular recanalization [Reference Jauch, Saver and Adams5].
Transcranial Doppler/duplex (TCD) ultrasound can display vessel stenoses, occlusions, and collateral flow, and thus help to understand the etiology and degree of compensation in AIS. However, TCD demands a good bony window in the patient, is quite investigator-dependent, time-consuming, and, compared to other means of intracranial vascular imaging, of less than optimal accuracy [Reference Feldmann, Wilterdink and Kosinski27], particularly with regard to the posterior circulation. Its use should not delay imaging in early AIS assessment. Secondary application of extracranial ultrasound after CT may help to detect stroke sources such as carotid stenosis or carotid/vertebral dissection, though. TCD has been suggested for enhancement of thrombolysis (sonothrombolysis), and initial studies were promising [Reference Alexandrov, Demchuk, Burgin, Robinson and Grotta28], but some authors reported cerebral hemorrhage when low frequencies were used [Reference Daffertshofer, Gass and Ringleb29], and since more evidence has been gathered in newer trials [Reference Alexandrov, Wojner and Grotta30,Reference Saqqur, Tsivgoulis and Molina31], TCD for sonothrombolysis cannot be routinely recommended.
Monitoring installed immediately on admission and remaining on the patient during acute stroke care and for at least the first 24 h has to include ECG for assessing cardiac rhythm and rate, blood pressure (BP) monitoring, and pulse oximetry for arterial oxygen saturation (SpO2). Temperature should be taken at least once on admission, and, if abnormal, periodically thereafter.
Acute Treatment of Early Ischemic Stroke
About 50–60% of patients with AIS might present with hypoxia or at least desaturation episodes (SpO2 <95% for >5 min) [Reference Sulter, Elting, Stewart, den Arend and De Keyser32], but the remaining patients are normoxemic. Still, the great majority of AIS patients receive supplementary oxygen in the prehospital phase and in the ER. Although it may seem intuitive that a patient with acute occlusion of a major brain artery, i.e. shutting off of oxygen from the respective brain region, should be supplied with extra oxygen, this is actually questionable. It is important to aim for tissue oxygenation and not for arbitrary levels of oxygen in the blood. Toxic levels of oxygen in patients that are not hypoxemic may cause tissue damage resulting from oxygen free radical formation, lipid peroxidation, and other mechanisms. Hyperoxia might also impede brain perfusion by a not completely understood process called hyperoxia-related cerebral vasoconstriction that may counterbalance and thus abolish the gain in blood oxygenation or might theoretically even lead to worsening of ischemia [Reference Iscoe and Fisher33–Reference Diringer35]. A large, quasi-randomized controlled trial resulted in no benefits from the application of 3 L oxygen within the first 24 h of AIS [Reference Ronning and Guldvog36]. It therefore seems reasonable to follow the guidelines at present [Reference Jauch, Saver and Adams5], i.e. to aim for no more than normoxemia and only administer oxygen noninvasively (e.g. by nasal pronges or face mask) to patients with an SpO2 below 95%. Respiration can often be supported by positioning the patient with the head of bed elevated to 15–30°. For some AIS patients who can tolerate it, the supine position might be preferrable [Reference Schwarz, Georgiadis, Aschoff and Schwab37,Reference Tyson and Nightingale38].
No clinical trial or prospective study has clarified the role of orotracheal intubation and early mechanical ventilation in AIS, although retrospective studies have suggested little benefit by these measures in the context of severe stroke [Reference Bushnell, Phillips-Bute and Laskowitz39,Reference Milhaud, Popp, Thouvenot, Heroum and Bonafe40]. However, if a patient has the perspective of potentially beneficial therapies such as endovascular recanalization, decompressive surgery, or critical care for treatable transient complications such as pneumonia, and the premorbid state and patient’s/family’s wishes do not preclude this, initiation of invasive ventilation should not be delayed if signs of respiratory decompensation or loss of airway protection are present.
Most AIS patients will present with high BP, either reflecting their underlying pathogenesis of a long-standing un- or undertreated arterial hypertension or an acute reaction of the body to raise cerebral perfusion, or both. In many cases, BP will decrease spontaneously within 90 min of onset. Since rapid reduction of BP may theoretically worsen penumbra perfusion in patients with compromised cerebral autoregulation or an (often unknown) hemodynamically relevant proximal artery stenosis, acute high BP should most often be tolerated. Despite several large studies on acute BP lowering in AIS, the data are very inconsistent and do not favor a particular regime or drug. Studies suggested a worse outcome if substantial BP variation was present in the first 24 h [Reference Delgado-Mederos, Ribo and Rovira41,Reference Stead, Gilmore and Vedula42]. A retrospective study in >10,000 AIS patients receiving IVT suggested that the best outcomes are associated with a systolic BP between 140 and 150 mmHg [Reference Ahmed, Wahlgren and Brainin43]. At present, it is advisable to tolerate acute high BP, if it does not exceed 220/120 mmHg. The latter is assumed to promote secondary hemorrhage, brain edema, or systemic end-organ damage. Furthermore, high BP must be lowered if IVT is planned. Some medical comorbidities, such as myocardial infarction, aortic dissection, or congestive heart failure, might also call for BP to be lowered and might result in a trade-off for optimal brain perfusion. BP management in these situations has not been clarified, and has to be handled individually and guided by close neurological monitoring to avoid deterioration. Oral premorbid antihypertensive drugs can be paused, since AIS patients may have impaired swallowing and the longer duration of pharmacologic action may compromise steering of BP in the acute phase. They may be restarted within the first few days, but when exactly is not clear and has to be decided individually. Table 11.6 shows the approach to hypertension in early AIS currently recommended in the guidelines [Reference Jauch, Saver and Adams5].
|Patient eligible for IVT, except if BP >185/110 mmHg||Labetalol 10–20 mg IV over 3 min, may repeat once; or|
|Nicardipine 5 mg/h IV, titrate up by 2.5 mg/h every 5–15 min, maximum 15 mg/h, adjust to maintain BP limits; or|
|Other appropriate agents (e.g. hydralazin, enalapril)|
|If BP cannot be lowered to <185/100 mmHg…||…do NOT administer recombinant tissue-type prothrombin activator (rtPA)!|
|Management of BP during and after recanalizing therapy to maintain BP <180/105 mmHg||Monitor BP every 15 min for 2 h from start of therapy, then every 30 min for 6 h, then every hour for 16 h|
|If systolic BP >180–230 or diastolic BP >105–120 mmHg||Labetalol 10 mg IV followed by continuous IV infusion 2–8 mg/min; or|
|Nicardipine 5 mg/h IV, titrate up to desired effect by 2.5 mg/h every 5–15 min, maximum 15 mg/h|
|If BP not controlled by above-named drugs or diastolic BP >140 mmHg||Consider IV sodium nitroprusside|
|Monitor and do not allow ICP to increase in a severe stroke|
Arterial hypotension is very rare in AIS and may indicate hypovolemia, heart failure, myocardial infarction, or sepsis [Reference Vemmos, Spengos and Tsivgoulis44], and should be treated adequately either by fluid administration or the use of vasopressors. Drug-induced hypertension for AIS outside these situations is not well supported by conclusive data. Hemodilution is not recommended.
There is ample evidence that acute hyperglycemia in AIS is associated with infarct growth and worse outcome [Reference Baird, Parsons and Phan45–Reference Ribo, Molina and Delgado47]. However, it is still unclear, despite considerable research effort, whether targeting glucose to a particular level in this acute phase is beneficial. GIST-UK, the only randomized trial on hyperglycemia treatment by insulin/potassium/glucose infusion yielded neutral results, was stopped prematurely, and had numerous methodological shortcomings [Reference Gray, Hildreth and Sandercock48]. In particular, the achieved difference in glucose levels was only ca. 10 mg/dL between the groups, quite possibly too little to achieve clinical significance. Hypoglycemia, although rare in AIS, may be equally if not more detrimental for the brain. Therefore, it is reasonable not to apply intensive glucose lowering, but to try to adjust the level between 140 and 180 mg/dL by subcutaneous insulin, and to treat hypoglycemia below 60 mg/dL vigorously, such as with a slow IV push of 25 mL of 50% dextrose [Reference Jauch, Saver and Adams5].
Anticoagulation and Antiaggregation
Past traditions of administering heparin or low molecular weight (LMW) heparinoids to AIS patients in the acute phase in order to promote perfusion have no convincing base of evidence. In addition, administration of direct thrombin inhibitors, other oral anticoagulants, IV glycoprotein IIb/IIIa antagonists, carries the risk of cerebral hemorrhage, is not established by data, and cannot be recommended, particularly not within the first 24 h after IVT. Oral aspirin within 48 h of stroke onset is recommended for most patients, but again not until 24 h after IVT [Reference Jauch, Saver and Adams5]. Withholding aspirin should be considered in patients with severe stroke if early decompressive surgery is planned.
It is of great importance, however, to clarify the coagulation status and the presence of premorbidly administered anticoagulants on admission, since this may influence decisions regading recanalization. In most cases, acute administration of anticoagulants or antagonization of previously taken anticoagulants will not be warranted in severe AIS.
Decision-Making According to Guidelines, Approval and Beyond
To reverse acute vessel occlusion, i.e. to recanalize, as soon as possible is not only a plausible principle, but was clearly associated with better outcome in several studies [Reference Nogueira, Liebeskind, Sung, Duckwiler and Smith49], and may be the paramount aim in acute stroke care. However, recanalization often has a limited time window and may eventually cause reperfusion trauma (relevant intracerebral hemorrhage, ICH). The slogan “TIME IS BRAIN” cannot be overemphasized. The time window, however, is highly individual and depends on the particular patient´s collateralization via adjacent vessels. In some patients, the penumbra (the brain area around the infarct core that is underperfused but potentially salvageable) may be lost within 2 h, while in others it may still be saved after 12 h or more. Therefore, larger stroke centers equipped with MRI or advanced CTA technology have turned to a more individual recanalization policy, as opposed to a rigid time window regime. Furthermore, since IVT will only lead to 30% recanalization in large-vessel (proximal main stem arteries) occlusion [Reference Rha and Saver50], endovascular interventions may be chosen instead or in addition. Also, although guideline-based recanalization by IVT include a great number of official contraindications, some of these (age over 80 years, diabetes, seizures, etc.) have been recognized to relate to design aspects of the approval studies for IVT, have been shown or suggested to be irrelevant by numerous other studies, and are ignored in many stroke centers, if the alternative (i.e. a large, severely disabling or life-threatening deficit) appears to outweigh the risks of treatment [Reference De Keyser, Gdovinova, Uyttenboogaart, Vroomen and Luijckx51] (Table 11.7). Furthermore, it was recently shown that patients with so-called minor strokes (NIHSS <4), but that still constitute an individually relevant deficit, should not be denied recanalization treatment. A considerable number of these will otherwise progress to a larger stroke and showed a favorable clinical course in several studies [Reference Kohrmann, Nowe and Huttner52,Reference Barber, Zhang, Demchuk, Hill and Buchan53]. Finally, IVT in stroke mimics (such as seizures or migraine attacks) seems to be quite safe and is only very rarely associated with symptomatic ICH [Reference Winkler, Fluri and Fuhr54–Reference Zinkstok, Engelter and Gensicke57].
|Inclusion criteria||Meaningful neurological deficit caused by ischemic stroke|
|Onset of symptoms <3 h (USA) before start of therapy|
|Age >18 years|
|Exclusion criteria||Frank hypodensities on CT being multilobar or involving >1/3 of the cerebral hemisphere|
|Prior stroke in previous three months|
|History or presence of intracranial hemorrhage|
|Suspicion of subarachnoid hemorrhage|
|Significant head trauma in previous three months|
|Intracranial neoplasm, arteriovenous malformation (AVM), or aneurysm|
|Recent intracranial or intraspinal surgery|
|Active internal bleeding|
|Acute bleeding diathesis, including but not limited to a platelet count of <100000/mm3|
|Heparin within the last 48 h raising aPTT above normal levels|
|Oral anticoagulation with INR >1.7 or PT >15 seconds|
|Current use of direct thrombin inhibitors or direct factor Xa inhibitors with elevated sensitive laboratory tests (such as aPTT, INR, platelet count, and ECT, TT, or appropriate factor Xa activity assays)|
|Highly elevated blood pressure (>185/100 mmHg)|
|Blood glucose concentration <50 mg/dL (2.7 mmol/L)|
|Relative exclusion criteriaa||Minor or rapidly improving symptoms|
|Age >80 years|
|Seizures at onset with postictal residual neurological symptoms|
|Major surgery or serious trauma in previous 14 days|
|Gastrointestinal or urinary tract hemorrhage in previous 21 days|
|Acute myocardial infarction in previous three months|
a Note: Recent evidence suggests that under some circumstances, with very careful consideration and weighting of risks and benefits, patients may receive IVT despite one or more of these contraindications! This policy should be restricted to experienced stroke centers. Adapted from [Reference Jauch, Saver and Adams5].
It has to be emphasized, however, that IVT by rtPA is only approved by the US FDA within 3 h and by the European EMA within 4.5 h of onset, according to the official indications and contraindications. All other approaches employing rtPA are off-label, individual attempts to save the patient and should ideally involve the patient’s or family’s written informed consent. If the patient is unable to provide this and the family or a legal proxy are not present, it is reasonable to proceed with the treatment regarded optimal in this emergency situation, but somebody bearing and constating witness to the physician´s decision is advisable.
The benefits of the current gold standard of recanalization were suggested by several observational studies before they were eventually confirmed in the NINDS milestone trial of 1995, which enrolled 624 patients and showed that AIS patients randomized to rtPA at a dose of 0.9mg/kg body weight within 3 h of stroke onset had an almost doubled chance of a favorable three-month outcome (complete or nearly complete recovery) when compared to placebo-treated patients . Several other degrees of deficits and measures of outcome were in favor of IVT, too. Interestingly, even patients with severe stroke (NIHSS >20, signs of brain edema and mass effect on CT), despite having an overall worse outcome, did better if they received IVT. The rate of (any, not necessarily symptomatic) ICH was 6.4% under rtPA and 0.6% under placebo, but mortality after two months was similar. These findings led to FDA and EMA approval of rtPA for IVT within 3 h of onset. The results were not only confirmed in several large subsequent trials, but also in the prospective community-based registry SITS-ISTR in almost 12,000 patients, i.e. in “real life” settings outside controlled trials [Reference Wahlgren, Ahmed and Davalos59]. Over the following 20 years, more pivotal trials have investigated extending IVT. After trials like ECASS I/II or ATLANTIS A/B had suggested that IVT up to 4.5 h from onset may be safe and beneficial, it was the second milestone trial, ECASS III in 2008, that provided proof. Of 418 AIS patients randomized to rtPA between 3 h and 4.5 h from onset, 52% had an excellent outcome (mRS 0–1) at three months compared to 45% of 403 placebo-treated patients (OR 1.34, p = 0.04) [Reference Hacke, Kaste and Bluhmki60]. Despite a rate of symptomatic ICH of 2.4% in rtPA-treated patients compared to 0.2% in those treated with placebo, mortality was higher in the latter group. In 2012, a meta-analysis of 12 trials on IVT (over 7000 patients, including results from the Third International Stroke Trial (IST-3), to date the largest RCT on IVT in 3035 patients), confirmed the previous findings, suggested benefits, even after up to 6 h, and in all age categories, and reinforced once more the importance of timely treatment [Reference Wardlaw, Murray and Berge61]. Based on data from these trials, the EMA approved rtPA for AIS up to 4.5 h after onset, while the FDA declined to do so, for reasons difficult to understand. US stroke experts from the AHA/ASA regard IVT within 4.5 h reasonable, but recommend following the ECASS III exclusion criteria (age >80 years, intake of oral anticoagulants, baseline NIHSS >25, CT hypodensities >1/3 of MCA territory, and the combination of previous stroke and diabetes mellitus) [Reference Jauch, Saver and Adams5]. Further extension of the time window, such as in wake-up stroke after selecting patients by stroke MRI (exclusion of advanced stroke by normal FLAIR sequence) has just been proven efficacious in the WAKE-UP trial [Reference Thomalla, Simonsen and Boutitie62].
In summary, there is no doubt that timely IVT works and improves the outcome of thousands of patients (Table 11.8). That benefits are achievable within a time window extending to 3 h does not mean that emergency physicians and neurointensivists may slow down. Treatment success remains clearly time dependent [Reference Lees, Ford and Muir63,Reference Lees, Bluhmki and von Kummer64]. The establishment of quick transport and access to recanalization by IVT in as many hospitals equipped with CT and basic stroke expertise as possible is certainly the most important action to improve stroke care world-wide. The employment of trained emergency personnel, the role of telemedicine, and even the option of CT-equipped stroke ambulances are currently under investigation to fulfill this goal. Emergency and critical care neurologists are key players in this scenario.
|Infuse 0.9 mg/kg rtPA (maximum 90 mg) over 60 minutes, with 10% of the dose given as a bolus over 1 min|
|Monitor patient on a stroke unit or a neurocritical care unit|
|Monitor BP and neurological state every 15 min for 2 h, then every 30 min for 6 h, then hourly for 24 h after start of rtPA infusion|
|Increase frequency of BP measurements if BP is >180/105 mmHg and administer antihypertensives to keep BP below this level|
|If the patient develops severe headache, acute hypertension, nausea or vomiting, neurological deterioration, assume ICH, stop rtPA infusion, obtain CT scan|
|Delay or avoid nasogastric tubes, indwelling bladder catheters, and central venous or arterial lines if the patient can be managed without|
|Obtain a follow-up CT or MRI 24 h after IVT before starting anticoagulants or antiplatelet drugs|
Application of rtPA via a well-secured IV access under continuous monitoring and controlled BP can be started in the ER and continued on the SU or the NCCU, or even in an ambulance, if transport to another center is decided. Close neurological assessment is paramount to detect deterioration as a possible sign of thrombolysis-related ICH. Another noteworthy side effect of IVT is orolingual angioedema, occurring in 1–5% of IVTs as swelling of lips, tongue, or oropharynx, usually contralateral to the affected hemisphere. Although most often mild and transient, it can rarely be life-threatening and may call for treatment interruption and antiallergic therapy (e.g. by ranitidine, methylprednisolone).
Intra-Arterial Thrombolyis, Mechanical Thrombectomy
Since large brain vessel occlusion (i.e. distal internal carotid artery (ICA), proximal middle cerebral artery (MCA, segments M1 or more than one M2), basilar artery (BA), or dominant vertebral artery (VA)) will respond to IVT in only 30% of the cases, alternative or additional endovascular local treatment (intra-arterial thrombolysis, IAT) may be chosen. Very different catheter devices, local therapies and adjunctive drug treatments have been tried over recent decades, and employed as IVT rescue strategies, replaced IVT in the case of contraindications, or combined with IVT (“bridging”) (for a review see [Reference Bosel, Hacke, Bendszus and Rohde65]). For bridging, prior IV rtPA is often reduced in dose, although it appears safe to use the full dose in bridging as well [Reference Nogueira, Yoo and Masrur66]. Despite impressive recanalization rates when combining IVT and IAT, a favorable outcome has only been achieved in up to 50% of cases [Reference Mazighi, Meseguer, Labreuche and Amarenco67]. Notwithstanding this lack of sufficient evidence of clinical efficacy, two devices, (MERCI and PENUMBRA) have been FDA approved. Today, mechanical thrombectomy (MT) using stent retrievers may be the most widespread and preferred technique, achieving recanalization rates of 80–100% [Reference Nogueira, Lutsep and Gupta68]. In February 2013, three randomized controlled trials (RCTs) on endovascular stroke treatment were published in the New England Journal of Medicine; none showed superiority over IVT [Reference Kidwell, Jahan and Gornbein69–Reference Broderick, Palesch and Demchuk71]. The largest of these, IMS3, was planned to randomize 900 patients to endovascular treatment (IAT or MT) added to IVT within 3 h versus IVT alone. After 656 patients the trial was stopped for futility and showed no difference in Modified Rankin Scale (mRS) or mortality at three months [Reference Broderick, Palesch and Demchuk71]. Since these trials hardly employed stent retrievers and were quite heterogeneous with regard to their design, their results were not regarded as settling the issue. Indeed, in 2014 the Dutch MR CLEAN trial was the first RCT to prove efficacy of endovascular AIS treatment. Of 500 patients with anterior circulation AIS, 190 were randomized for MT by stent retrievers within 6 h of onset in addition to IVT, 32.6% of which reached an mRS of 0–2 at 90 days compared to 19.1% in the IVT-only group [Reference Berkhemer, Fransen and Beumer72]. Soon, two more highly positive RCTs on MT with stent retrievers were published: ESCAPE, demonstrating favorable long-term outcome in 53% (IVT+) IAT patients versus 29.3% IVT patients out of 316 anterior circulation AIS [Reference Goyal, Demchuk and Menon73], and EXTEND-IA, showing highly significant benefits of IAT combined with IVT in only 70 AIS patients as compared to IVT only, i.e. early neurological improvement and better reperfusion in 80% and 100% versus 37% and 37%, respectively [Reference Campbell, Mitchell and Kleinig74]. Both trials employed stent retrievers, multimodal CT imaging, and short treatment windows. More RCTs from independent groups all over the world that were stopped early or completed thereafter, all showed consistently that the combination of MT with IVT was more efficacious than IVT alone for AIS caused by LVO [Reference Saver, Goyal and Bonafe75–Reference Saver, Goyal and van der Lugt77]. These results substantially affected acute stroke care and influenced guidelines [Reference Powers, Derdeyn and Biller78]. The newest development at the time of writing is that selected stroke patients with LVO and good collaterals can even be selected neuroradiologically to benefit from MT as late as 16–24 h from stroke onset [Reference Nogueira, Jadhav and Haussen79]. Currently, MT for mild and more severe strokes and patients with poorer imaging status is being investigated. A number of questions remain to be answered, such as how to best manage the patients peri- and postprocedurally. At present, key elements for success appear to be the use of stent retrievers, detection of salvageable brain tissue by rapid and simply interpretable imaging, and a lean flow of action.
Neurointensivists, neuroanesthesists, and emergency physicians may be involved in stabilizing the patient before, during, and after endovascular stroke. A standardized protocol for the procedure can help to save valuable time [Reference Herrmann, Hug and Bosel80]. It appears reasonable, although not sufficiently backed up by studies, to keep oxygenation and circulation parameters within certain limits during the procedure. Examples from the Heidelberg stroke center are SpO2 >95%, ETCO2 35–45 mmHg, SBP 140–160 mmHg. Great variations of blood pressure and especially hypotension should be avoided [Reference Delgado-Mederos, Ribo and Rovira41,Reference Stead, Gilmore and Vedula42,Reference Reich, Hossain and Krol81]. In addition, hyperventilation associated with hypocarbia may theoretically lead to cerebral vasoconstriction and hypoperfusion of the penumbra and may be deleterious. Both events may occur during intubation and mechanical ventilation [Reference Takahashi, Brambrink and Aziz82]. Although most interventionalists prefer general anesthesia and intubation for the procedure [Reference McDonagh, Olson and Kalia83], a number of retrospective studies suggest that the nonintubated state under so-called “conscious sedation” may be advantageous and even result in better outcomes [Reference Abou-Chebl, Lin and Hussain84–Reference Davis, Menon and Baghirzada86]. This, however, was contradicted by three single-center randomized trials comparing general anesthesia with conscious sedation. In all three, general anesthesia was not inferior, and in some regards better, compared to conscious sedation [Reference Schonenberger, Uhlmann and Hacke87–Reference Simonsen, Yoo and Sorensen89]. Many other open questions on peri- and postinterventional care exist, such as those on adjunctive anticoagulation, temperature management, timing of extubation, timing of control CT scan, etc. At present, it may be reasonable to avoid adjunctive anticoagulants, to keep the patient normothermic, apply immediate close NCCU monitoring, strive for early extubation, and perform a CT scan 12 h after the procedure or earlier, if the patient does not wake up for extubation or displays neurological deterioration [Reference Bosel90].
General Aspects of Ischemic Stroke Management in the NCCU
Since high-quality data on many aspects of general critical care specific for ischemic stroke are scarce, the reader will be referred to general stroke guidelines [Reference Jauch, Saver and Adams5,91] and general intensive care guidelines (on specific topics of ICU or NCCU care, as published by the SCCM, the ESICM, and the NCS) throughout this part of the chapter. Careful adaptation of these may often be appropriate for the NCCU patient with ischemic stroke as well.
Impaired level of consciousness, decreased respiratory drive, loss of protective reflexes and dysphagia may lead to life-threatening respiratory situations in patients with severe AIS on admission or during the later course. In the past, numerous studies suggested that AIS patients so severely affected that they required invasive airway securing and mechanical ventilation [Reference Steiner, Mendoza and De Georgia3,Reference el-Ad, Bornstein, Fuchs and Korczyn4,Reference Berrouschot, Rossler, Koster and Schneider92–Reference Santoli, De Jonghe and Hayon95] would have mortality rates between 40 and 80%. The relative value of intubating AIS patients, considering the high use of ICU resources and the bad prognosis despite ICU care, was challenged by these reports. However, these studies may have been influenced by a nihilistic or self-fulfilling prophecy in the absence of a proven effective treatment option. Advanced options for therapies of severe AIS, including the potential perspective of decompressive surgery, support that life-saving intubation and initiation of mechanical ventilation should not be withheld or delayed. Certainly, this should take into account the wishes of the patient or the family and the overall clinical situation, but initating invasive ventilation in an acute and possibly unclear emergency situation when such statements cannot be obtained does not constitute an ethical obstacle to later withdrawing treatment efforts, if appropriate.
The need for mechanical ventilation in the large hemispherical infarction (LHI) patient may often be driven by the need for airway protection in a conscious adult, and may therefore be different from other populations. One prospective observational study addressed mechanical ventilation and intubation in LHI [Reference Milhaud, Popp, Thouvenot, Heroum and Bonafe40], reporting a GCS score of <10 or respiratory failure as indications for intubation. In another prospective study, 54 of 218 AIS patients required mechanical ventilation; of these, 90% were intubated due to neurological deterioration and 10% due to cardiopulmonary compromise [Reference Berrouschot, Rossler, Koster and Schneider92]. Independent predictors of mechanical ventilation (MV) were history of hypertension and infarct size >2/3 of MCA territory. In AIS patients with – at times even relatively small – vertebrobasilar strokes, intubation is frequently necessary if the brainstem is involved and decline of level of consciousness, dysphagia, or loss of protective reflexes ensue. The decision to intubate should be made using features such as: (a) a GCS ≤8 or progressive decline in level of consciousness, (b) clinical or monitoring signs of respiratory failure, (c) loss of protective reflexes, (d) signs of increased intracranial presure, (e) infarct size >2/3 of MCA territory on imaging [Reference el-Ad, Bornstein, Fuchs and Korczyn4], (f) coexistence of pulmonary edema or pneumonia, or (g) imminent surgical or invasive procedure.
Principally, early extubation should be a primary goal. Extubation can be difficult, as impaired level of consciousness and a high prevalence of dysphagia may lead to extubation failure and reintubation, which is associated with increased morbidity and mortality in ICU patients [Reference Esteban, Alia and Gordo96]. NCCU patients often experience postextubation dysphagia [Reference Langmore97], and reintubation rates can be as high as 35% [Reference Vallverdu, Calaf and Subirana98,Reference Navalesi, Frigerio and Moretti99]. Classical predictors of successful extubation from general critical care are unreliable in the brain-injured ICU patient [Reference Ko, Ramos and Chalela100,Reference Coplin, Pierson, Cooley, Newell and Rubenfeld101] and this certainly applies to most AIS patients in the NCCU as well. For instance, cooperation is often not achievable simply due to aphasia or apraxia, and dysphagia is much more frequent. As such, classical extubation triggers can only be used for cautious orientation in LHI patients. Only one recent retrospective study in 47 intubated MCA stroke patients suggested that a composite GCS score ≥8 trends towards extubation success (with a mean eye score of 4 in those who could be extubated versus 2.5 in those who could not) [Reference Wendell, Raser, Kasner and Park102]. Prospective studies focused on extubation success in AIS NCCU patients do not exist. A recent prospective study, PRINCIPLE, only reported in abstract so far, investigated predictors of reintubation in 93 critically ill AIS patients planned for extubation. Of these, 36% needed reintubation, and the predictors were reduced level of consciousness for initial intubation, a higher Airway Care Score, episodes of raised ICP, length of NCCU stay, and need for antibiotic treatment. Obviously, extubation must not be attempted if sufficient respiratory and airway protection criteria [Reference Boles, Bion and Connors103] are not present, but if they are, it often can be achieved even though cooperation is not established. On the other hand, failure to detect dysphagia might result in unnecessary reintubation. It appears reasonable to attempt extubation after successful spontaneous breathing trials, in the absence of relevant oropharyngeal saliva collections and absence of relevant demand for suctioning, together with the presence of a cough reflex and tube intolerance, if the patient is free of analgesia and sedation, even if communication and cooperation cannot be established. Stand-by reintubation measures should be provided, and the patient remain under monitored observation for at least 24 h postextubation.
Tracheostomy is frequently necessary in the ICU patient if timely extubation is not feasible. While the procedure has to be applied to ca. 10–15% in the general ICU, the rate is 20–30% in the NCCU [Reference Kurtz, Fitts and Sumer104,Reference Pelosi, Ferguson and Frutos-Vivar105]. Tracheostomy and particularly early tracheostomy, reportedly beneficial in selected ICU patients [Reference Griffiths, Barber, Morgan and Young106], has seldom been studied in AIS patients specifically [Reference Bosel, Schiller, Hacke and Steiner107,Reference Bosel108]. A retrospective study in ICU stroke patients (including AIS) surviving ICU care after prolonged ventilation and tracheostomy suggested good outcomes in about 25% of tracheostomized patients, as well as reductions in ventilation duration, ICU stay, and costs from earlier tracheostomy [Reference Rabinstein and Wijdicks109]. A more recent randomized trial on early tracheostomy in mixed cerebrovascular ICU patients including 20 patients with large ischemic hemispheric stroke did not confirm the latter two advantages but demonstrated safety, feasibility, and reduction in the need for sedation [Reference Bosel, Schiller and Hook110]. Predictors for tracheostomy need in patients with severe AIS have not been defined and there is currently insufficient evidence on potential outcome benefits of tracheostomy in general or early tracheostomy in particular. As there is some evidence and also experience-based reasons to assume tracheostomy is a relatively safe and feasible bedside-procedure in the NCCU [Reference Bosel, Schiller and Hook110,Reference Seder, Lee and Rahman111], this should be done as early as the need for long-term airway protection or mechanical ventilation becomes clear to the treating physican. This remains a clinical, individual decision and general customs for tracheostomy in ICU patients should be applied. It appears reasonable to consider tracheostomy in AIS patients failing extubation or if extubation is not feasible by 7–14 days after intubation.
Modern ventilation of the ICU patient follows lung-protective principles. One of these is the application of low tidal volume at higher respiratory frequencies. The associated benefits were first shown in acute respiratory distress syndrome (ARDS) patients and then extended to other ICU populations [Reference Serpa Neto, Cardoso and Manetta112]. Another principle is that of the “open lung,” meaning to keep alveoli open by principal application of positive end expiratory pressure (PEEP) and that of PEEP escalation in the treatment of ARDS [Reference Briel, Meade and Mercat113]. Superiority of any specific mode of mechanical ventilation has not been established in the ICU, so far. Although the above-named principles of ventilation have not been studied sufficiently in NCCU-dependent AIS patients, it appears reasonable to apply them, until such data support other modes. Goals of ventilation should be optimal arterial oxygenation (arterial partial pressure of oxygen (PaO2) >70 mmHg) and normalization of the arterial partial pressure of carbon dioxide (PaCO2) to between 35 and 40 mmHg, as a drop in the latter may lead to cerebral vasoconstriction and risk of secondary ischemia, while a rise may cause cerebral vasodilation and a subsequent increase in ICP. It appears desirable to strive for early adoption of patient-controlled modes of ventilation to achieve training of respiratory muscles, and to apply standardized respirator weaning as soon as the patient allows this.
Two small prospective observational studies have addressed ventilation settings in LHI patients. The fear of increased venous return resulting in increased ICP when PEEP escalation is used to treat LHI patients was first examined in a study in 13 LHI patients equipped with ICP neuromonitoring. Increasing PEEP to 12 mmHg did not lead to relevant rises in ICP [Reference Georgiadis, Schwarz, Baumgartner, Veltkamp and Schwab114]. Further, intensifying ventilation using alterations of the inspiration:expiration (I:E) ratio from 1:2 to 1:1 in 13 LHI patients did not influence ICP or cerebral perfusion pressure (CPP) [Reference Georgiadis, Schwarz, Kollmar, Baumgartner and Schwab115]. Thus, these forms of ventilation escalation to improve oxygenation may be safe in LHI patients, provided neuromonitoring (ICP and CPP) is in place. In the only prospective study on mechanical ventilation in 37 patients with AIS, including LHI, subjects were ventilated using a volume-controlled mode and an I:E ratio of 1:2, the PaO2 was adjusted to >90 mmHg and the PaCO2 to about 35 mmHg. In this population, no patient received hemicraniectomy, and mortality was 70%, irrespective of the cause for intubation (“respiratory” versus “neurological”). No further focus on the method of ventilation or subgroup analysis was chosen in this study [Reference Milhaud, Popp, Thouvenot, Heroum and Bonafe40].
Until new data suggests otherwise, patients with severe AIS should be ventilated according to general principles of modern ICU ventilation with the goals of normoxemia and avoidance of hypo- and hypercapnia. Escalation of PEEP or alteration of I:E ratio can be considered to achieve optimal oxygenation, preferrably if ICP/CPP monitoring is in place.
Analgesia and Sedation
Patients with large AIS, in the acute phase of their disease, often need analgesia and sedation to achieve freedom from pain, anxiety, and agitation. Addtionally, sedation and analgesia may facilitate medical goals such as lowering ICP, enabling procedures and operations, or terminating seizures. The way to optimally use analgesia and sedation has rarely been studied in NCCU patients, including those with AIS. Although common analgesics and sedatives such as different opioids, midazolam, propofol, ketamine, etc. have been subject to small studies on brain-injured patients (mainly traumatic brain injury (TBI) or subarachnoid hemorrhage (SAH)), but this was largely limited to physiologic rather than outcome effects. There are currently no data to allow for preference of any analgesic or sedative agent over another in neurocritical care [Reference Teitelbaum, Ayoub and Skrobik116,Reference Reade and Finfer117]. Likewise, the application of sedation and pain scores, sedation and analgesia protocols [Reference Egerod, Jensen, Herling and Welling118,Reference Karabinis, Mandragos and Stergiopoulos119], or sedation monitoring devices such as bispectral index (BIS) have been addressed in small studies on brain-injured ICU patients [Reference Olson, Thoyre and Auyong120,Reference Olson, Thoyre, Peterson and Graffagnino121], but not in AIS patients, specifically. Daily wake-up trials were reported to be beneficial for reduction of ventilation duration [Reference Kress, Pohlman, O’Connor and Hall122] and outcome [Reference Girard, Kress and Fuchs123,Reference Hooper and Girard124] in earlier ICU studies. A large recent RCT failed to confirm a benefit associated with daily wake-up protocols [Reference Mehta, Burry and Cook125]. Furthermore, sedation interruption in NCCU populations (TBI and SAH) was associated with potentially negative effects, such as transient rises in ICP and stress hormone levels [Reference Skoglund, Enblad and Marklund126,Reference Skoglund, Enblad, Hillered and Marklund127] and cerebral deoxygenation [Reference Helbok, Kurtz and Schmidt128], and this might apply to AIS patients as well.
Aiming for the lowest level of analgesia and sedation that still provides systemic and cerebral hemodynamic stability, as well as patient comfort, seems desirable. Overall, the goal should be to free the patient from sedation as soon as appropriate. Neuromonitoring of at least ICP and CPP is recommended to guide sedation. Daily wake-up trials should be abandoned or postponed at clinical or monitoring signs of physiological compromise or discomfort. Intensivists should use the agents customary in their units, choosing them according to patient comorbidities and agent-specific side effects, and pay particular attention to avoidance of hypotension.
Dysphagia as a result of stroke lesions compromising the swallowing act at different stages affects 30–50% of stroke patients in the acute phase. The strongest impairment is caused by brainstem stroke. Screening for dysphagia has been reported to decrease pneumonia in the general stroke population [Reference Yeh, Huang and Wang129]. Dysphagia screening tests such as the gugging swallowing test (GUSS) have been studied and found useful in AIS patients, but patients with large or multiple strokes or rapid decline in level of consciousness were not included [Reference Trapl, Enderle and Nowotny130]. The swallowing provocation test (SPT), testing the involuntary part of swallowing by means of a thin oropharyngeal catheter, might overcome problems with vigilance or cooperation [Reference Warnecke, Teismann and Meimann131]. After initial screening, dysphagia can be confirmed and differentiated by endoscopic swallowing tests that do not necessarily demand patient cooperation. In particular, the fiberoptic endoscopic evaluation of swallowing (FEES) can be done in severely affected and uncooperative patients, and it has been found reliable and predictive in studies on acute stroke patients [Reference Dziewas, Warnecke and Olenberg132,Reference Warnecke, Teismann and Oelenberg133]. A recent study of over 300,000 stroke patients found that less than one-third received dysphagia screen and 1 in 17 experienced a hospital-associated pneumonia [Reference Masrur, Smith and Saver134].
Studies on the best timing for nasogastric tube (NGT) placement [Reference Dennis, Lewis and Warlow135] in critically ill AIS patients are lacking. Similarly, predictors of the need for percutaneous endoscopic gastrostomy (PEG) tube placement have only been studied in mixed ischemic stroke populations. Two retrospective analyses have found a high NIHSS score to be most predictive of PEG need [Reference Kumar, Langmore and Goddeau136,Reference Alshekhlee, Ranawat and Syed137]. In a recent Cochrane analysis of the general acute and subacute stroke population, NGT and PEG did not differ in terms of case fatality, but PEG was more secure and resulted in fewer treatment failures and reduced gastrointestinal bleeding, as well as higher feed delivery [Reference Geeganage, Beavan, Ellender and Bath138]. In terms of gastric ulcer prophylaxis, recent data support ranitidine use in reduction of nosocomial pneumonia versus proton pump inhibitors.
With this lack of studies focusing on food intake, delivery, and gastrointestinal function in NCCU patients with AIS, it may be best to carefully transfer insights from general ICU care. It is often possible to assess swallowing by standard screening tests in the very early phase (days 1 and 2; patients might still be in the emergency room or the stroke unit) to guide NGT placement. Once level of consciousness starts to decline, however, the patient should be kept nil by mouth and the NGT be placed at a very low threshold to avoid aspiration. The NGT can then be used for feeding over the next few weeks of NCCU treatment. During this time, gastrointestinal transport stimulation and gastric ulcer prophylaxis following principles from general critical care can be applied. In the later phase of the disease, i.e. after termination of sedation and weaning from the respirator, swallowing capacity can be reassessed, preferrably by use of an endoscopic method such as the FEES and the results incorporated into the decision to place a PEG.
Fluid Status and Nutrition
AIS patients in the NCCU should be kept euvolemic, since hypovolemia may compromise cerebral perfusion by hypotension and high blood viscosity, while hypervolemia may promote brain edema. The average demand for an adult patient is 2–3 L/day (1 mL/kg/h), but individual factors such as higher perspiration under invasive ventilation and fever, as well as cardiac and renal function have to be taken into account. After decades of ICU controversy regarding colloidals and crystalloidals in volume management, more recent studies have cast doubt on the benefits of colloids in many situations and populations. ALIAS, a large RCT on albumin versus crystalloids in AIS was stopped early for futility after 841 patients, since no outcome benefits in either group were found [Reference Ginsberg, Palesch and Hill139]. Hypotonic fluids may promote edema and should be avoided. In essence, crystalloids (normal saline) should be chosen for fluid management of the AIS patient in most situations.
In terms of nutrition, one study addressed the energy demand of 21 LHI patients in a group of 35 stroke patients [Reference Bardutzky, Georgiadis, Kollmar, Schwarz and Schwab140]. In these sedated and ventilated patients, the total energy expenditure (TEE) was assessed by indirect calometry during the first five days after admission. A strong correlation was found between the TEE and the basal energy expenditure (BEE), as estimated by the Harris–Benedict equation. The study showed a lower energy expenditure in these LHI patients than in other ICU populations.
The guiding principle in general ICU nutrition is towards a preference for enteral over parenteral nutrition. A recent meta-analysis demonstrated benefits of early (<24 h) initiation of enteral nutrition [Reference Doig, Heighes, Simpson, Sweetman and Davies141]. Nutrition should be tailored to meeting caloric demand while avoiding overfeeding and hyperglycemia in AIS patients, using the Harris–Benedict equation to predict basal energy demand. However, this formula does not provide information on the composition or delivery method of nutrition. It is probably advisable to follow general ICU principles in AIS patients, i.e. to: (a) calculate energy (caloric) demand and standardize caloric requirements based on height, weight, age, and sex, (b) balance nutrition to represent the components carbohydrate, protein, and fat (taking into account “nutritional” components of some infused drugs) and supplement, (c) start nutrition on day 2 from admission, (d) use enteral over parenteral nutrition whenever possible.
Both hyperglycemia and hypoglycemia are associated with increased morbidity and mortality in severe AIS and many other neurocritical care conditions. So far, it remains unclear whether systemic glucose itself is the decisive pathophysiologic factor in these observations or just an indicator of other compromising mechanisms. Likewise, the impact of controlling glucose and the best method to do so have remained controversial in stroke patients [Reference Gray, Hildreth and Sandercock48] in the brain-injured ICU population. A recent large retrospective analysis of more than 1900 mixed NCCU patients showed no benefits of tight glucose control, but rather a higher rate of hypoglycemia and higher mortality [Reference Graffagnino, Gurram, Kolls and Olson142]. Furthermore, a recent systematic review and meta-analysis on 16 RCTs in more than 1200 mixed NICU patients found no mortality benefits of intensive insulin therapy, but the association with a higher risk of hypoglycemia, particularly jeopardizing in the ischemic brain. However, very loose glycemic control was also associated with worse neurologic recovery. The authors concluded that intermediate glucose control (insulin therapy aiming for 140–180 mg/dL) may be most appropriate for this patient population [Reference Kramer, Roberts and Zygun143] and is also recommended in current stroke guidelines[Reference Jauch, Saver and Adams5].
Anemia is associated with worse outcome in ischemic stroke, if it is present in either the acute [Reference Tanne, Molshatzki and Merzeliak144] or the subacute [Reference Kellert, Martin and Sykora145] phase. Anemia, found in the majority of ICU patients from the third day of admission, has also been associated with worse outcome in brain diseases such as SAH, TBI, and ICH (for a review see [Reference Kramer and Zygun146]). A very recent retrospective study on critically ill AIS patients revealed that almost all of these acquire marked anemia during their NCCU course, which is then associated with longer ICU stay and duration of ventilation [Reference Kellert, Schrader, Ringleb, Steiner and Bosel147]. However, the optimal hemoglobin level in critically ill AIS patients is unclear, as is the optimal red blood cell (RBC) transfusion policy. A recent systemic review and meta-analysis found six RCTs concerning hemoglobin levels and transfusion in the NICU and came to the conclusion that insufficient evidence exists to currently recommend either a restrictive or a liberal transfusion strategy in these patients [Reference Desjardins, Turgeon and Tremblay148]. However, AIS patients were not the subject of the trials analyzed.
Theoretically, optimizing the oxygen carrying capacity should play a decisive role in ischemia and oligemia. The physiological benefits of RBC transfusion suggested in neuromonitoring studies in TBI and SAH patients [Reference Zygun, Nortje and Hutchinson149,Reference Smith, Stiefel and Magge150] should be assumed for LHI patients as well. It is thus questionable if 7 g/dL hemoglobin, (a widely accepted level in general ICU patients), is also optimal for LHI patients. However, transfusion was more often found to be associated with worse outcome in neurocritical care patients, either by being an indicator of higher disease severity or indeed causing harm [Reference Kramer and Zygun146]. In anemic NCCU patients with AIS, RBC transfusion did not improve the situation [Reference Kellert, Schrader, Ringleb, Steiner and Bosel147]. Until more research clarifies optimal (individual?) hemoglobin levels and transfusion strategies, it is probably reasonable to use RBC transfusion below a hemoglobin level of 7 g/dL or if systemic or cerebral monitoring – if available – suggests a compromised arteriovenous oxygen extraction. RBC transfusion should also be triggered by specific situations such as planned surgery, hemodynamic status, and bleeding. Anemia or the progression thereof should be avoided by reducing blood sampling to the minimum necessary, treating infection and renal disease early, and preventing volume overload and promoting appropriate diuresis, etc.