Background

Patients with acute stroke often have breathing disturbances leading to hypoxaemia (Sulter et al., 2000; Roffe et al., 2003). Severe hypoxaemia can result in cerebral hypoxia, which can exacerbate the ischaemic brain injury and cause new brain injury. Therefore, it is important to monitor breathing and oxygenation and prevent or correct hypoxaemia promptly.

Common causes of hypoxaemia during acute stroke include partial airway obstruction, hypoventilation, aspiration, atelectasis, and pneumonia. Patients with decreased consciousness or brainstem dysfunction are at increased risk of airway compromise because of impaired oropharyngeal mobility and loss of protective reflexes. In addition to airway obstruction, some acute stroke patients develop irregular breathing patterns, such as Cheyne–Stokes respirations, that further compromise blood oxygenation. Some acute stroke patients have dysphagia, placing them at risk for aspiration and further hypoxaemia. Hypoxaemia is also common during seizures, which can be a complication of acute stroke. Also, head of bed (HOB) position has been suggested to affect oxygenation in acute stroke patients.

Evidence

A total of at least 11 randomized trials have studied normobaric oxygen supplementation to non-hypoxaemic acute stroke patients, without strong evidence of beneficial effects (Ding et al., 2018). The largest randomized trial of prophylactic low-dose oxygen supplementation randomized 8003 acute stroke patients to one of three groups: control, nocturnal oxygen only, or continuous oxygen (Roffe et al., 2017). Nasal oxygen was administered within 24 hours of hospital admission at 2 L/min when the oxygen saturation was above 93% and at 3 L/min when it was at or below 93%. Protocol treatment continued for 72 hours. The primary functional clinical outcome, ordinal modified Rankin Scale (mRS) of global disability at 3 months, was similar in the three treatment groups, with odds ratio (OR) for better outcome of 0.97 (95% CI: 0.89–1.05; p = 0.47) for any oxygen versus control, and OR of 1.03 (95% CI: 0.93–1.13; p =  0.61) for continuous vs nocturnal oxygen. A meta-analysis adding to this trial the 4 additional small trials with mRS distribution data also showed no beneficial effect of continuous oxygen supplementation, with a mean difference in mRS scores of 0.00 (95% CI: –0.15–0.14; p = 0.97).

Similarly, trials have not demonstrated a benefit of hyperbaric oxygen therapy in acute ischaemic stroke, although studies have been small in size. Among 11 randomized trials enrolling a total of 705 patients, mortality at 6 months did not differ with hyperbaric oxygen therapy, with risk ratio (RR) of 0.97 (95% CI: 0.34–2.75; p = 0.96) (Bennett et al., 2014). Functional outcome scales varied widely across trials, precluding pooled analysis, but the available data, though not excluding a therapeutic effect, were not strongly suggestive.

One early systematic review of HOB positioning in early stroke patients included 4 studies enrolling 183 patients. Different endpoints prevented formal meta-analysis, but individual study results indicated that head position does not affect blood oxygenation in stroke patients without respiratory disorders, while in patients with respiratory disorders HOB elevation is likely to improve oxygenation (Tyson et al., 2004). A large, international, cluster-randomized crossover trial of HOB position with 11,093 acute stroke (85% ischaemic) patients tested horizontal (supine) versus a 30° reclining position maintained for 24 hours (Anderson et al., 2017) (also see section on HOB position in this chapter). This study found no difference in oxygenation status or the primary functional outcome between the two treatment groups.

Comment

While management of non-hypoxaemic acute stroke patients may consist of only monitoring, hypoxaemia should be corrected promptly. Continuous pulse oximetry is a standard, simple, and reliable method to monitor blood oxygenation (Sulter et al., 2000; Kim et al., 2017). In addition to oximetry, periodic assessment of breathing pattern, breathing rate, and breath sounds can detect impending respiratory insufficiency. Blood gas analysis and end tidal CO₂ monitoring help to confirm and define respiratory insufficiency and to guide further treatment. Any effect of HOB position on oxygenation should be noted and used to optimize oxygenation.

Blood oxygen saturation should be maintained above 94%. Noninvasive measures should be considered first, such as nasal cannula, Venturi mask, non-rebreather mask, bilevel positive airway pressure, and continuous positive airway pressure.

Endotracheal intubation with mechanical ventilation should be reserved for patients who do not respond sufficiently to the noninvasive measures or are in need of airway protection from aspiration. Mechanical ventilation can also be useful in the management of stroke patients with malignant cerebral oedema (see Chapter 11). In addition to supplemental oxygen and ventilation assistance, the causes of hypoxaemia should be investigated and treated urgently. The timing of initiation of mechanical ventilation is important. Once a clear indication for mechanical ventilation is identified, urgent endotracheal intubation should be done in order to enhance oxygenation and minimize complications (Steiner et al., 1997).

Background

Head positioning can potentially modify acute ischaemic stroke outcome through multiple mechanisms. During acute ischaemic stroke, cerebral autoregulation is impaired in the vicinity of the ischaemic zone, and changes in blood pressure (see blood pressure section in this chapter) as well as HOB position may affect cerebral perfusion and the ultimate fate of the brain region at risk (cerebral penumbra). Also, in patients with large strokes and significant brain oedema, the effect of intracranial pressure on cerebral perfusion needs to be considered when altering HOB position. In addition, head position may affect saliva pooling and aspiration risk. In many clinical situations, the risks and benefits of HOB recommendations need to be balanced with limited data.

Evidence

Multiple small studies of head positioning in acute stroke found a decrease in middle cerebral artery blood flow velocity with HOB elevation from 0° to 30° or 0° to 45° (Schwarz et al., 2002; Wojner-Alexandrov et al., 2005; Aries et al., 2013). In one study, these changes were similar between the stroke and the unaffected hemisphere, and between stroke patients and healthy controls (Aries et al., 2013). One small study measured cerebral perfusion pressure (CPP) in acute stroke patients and found a decrease with head position elevation and associated falls in mean arterial pressure and intracranial pressure (Schwarz et al., 2002).

Multiple studies measured frontal lobe blood flow in acute stroke patients with a transcranial optical spectroscopy method (Durduran et al., 2009; Aries et al., 2013; Favilla et al., 2014) and found a decrease with HOB elevation. However, a considerable heterogeneity in response was noted (Aries et al., 2013).

In a large, international, cluster-randomized crossover trial of head positioning, 11,093 acute stroke (85% ischaemic) patients were allocated to either lying flat or sitting up with at least 30° elevation soon after hospital admission and maintained for 24 hours (Anderson et al., 2017). Patients in this study had relatively mild strokes with median baseline National Institutes of Health Stroke Scale (NIHSS) score of 4. The median time from stroke onset to initiation of head positioning was 14 hours. There was no significant difference in the primary functional outcome of shift of disability level on the modified Rankin Scale at 90 days, lying flat versus sitting-up: OR 1.01 (95% CI: 0.92–1.10; p = 0.84). Similarly, no differences were seen in the rate of functional independence (mRS: 0–2) at 90 days, lying flat versus sitting-up: 38.9% versus 39.7%, OR 0.94 (95% CI: 0.85–1.02; p = 0.25).

Comment

The numerous small studies of cerebral haemodynamics in acute stroke suggest potentially important effects of head positioning on regional cerebral blood flow (CBF) in some patients, and they also indicate considerable variability in responses to head position. These findings, together with the neutral result from the large randomized trial (Anderson et al., 2017) and the heterogeneous nature of ischaemic stroke, suggest that HOB position does not have a major effect upon most patients. The possibility remains that in a small subset of exquisitely collateral-dependent hyperacute stroke patients, head positioning importantly alters course, but this requires rigorous studies to confirm or discount.

Background

Normally, CBF is maintained relatively constant during arterial blood pressure fluctuations by a physiological system termed cerebral autoregulation. Cerebral autoregulation occurs as the cerebral blood vessels dilate (creating decreased resistance to blood flow) in response to falls in pressure and constrict (creating increased resistance to blood flow) in response to increases in pressure. However, cerebral autoregulation has upper and lower blood pressure limits. Below certain blood pressures the cerebral autoregulation is exceeded and CBF falls, with possible brain ischaemia. Also, above certain blood pressures the cerebral autoregulation is exceeded and the CBF rises, with possible brain oedema. The limits of cerebral autoregulation are related to baseline blood pressures (Strandgaard et al., 1973; Schmidt et al., 1990; Donnelly et al., 2016). The lower limits of cerebral autoregulation are approximately 60–85 mm Hg mean pressure in normotensive individuals and approximately 110–120 mm Hg mean pressure in uncontrolled hypertensive individuals.

In the vulnerable region of an acute ischaemic stroke, the cerebral penumbra, cerebral autoregulation is impaired and the CBF is directly related to changes in blood pressure (Dirnagl and Pulsinelli, 1990; Donnelly et al., 2016). Therefore, blood pressure reductions during acute ischaemic stroke may decrease the CBF and extend the ischaemic injury in the cerebral penumbra. During acute ischaemic stroke, the blood pressure is naturally elevated to various degrees above usual. Observational acute ischaemic stroke studies have found a U-shaped relation between the acute blood pressure and clinical outcomes (Castillo et al., 2004; Ahmed et al., 2009). Clinical outcomes were worse at the extremes of blood pressures than at the in-between ranges. In one large study, a systolic blood pressure range of 141–150 mm Hg was associated with best functional outcome (Ahmed et al., 2009). Blood pressure management during the acute period of ischaemic stroke has been the focus of multiple clinical trials.

Evidence

Blood Pressure Lowering

A Cochrane systematic review identified 10 randomized trials of blood pressure lowering in acute ischaemic stroke, enrolling 11,238 patients (Bath and Krishnan, 2014). Tested agents included angiotensin-converting enzyme (ACE) inhibitors, angiotensin receptor antagonists, beta-blockers, nitric oxide donors, and other agents used at physician discretion. No net effect on death by end of trial follow-up was noted, 6.0% versus 6.3%, OR 0.95 (95% CI: 0.78–1.16). In addition, in the 8 randomized trials reporting death or dependency at end of trial, no effect was observed between blood pressure lowering versus control, 41.1% versus 41.0%, OR 1.00 (95% CI: 0.92–1.08) (Figure 5.1). A subsequent systematic review identifying additional studies, with 13 trials enrolling 12,703 patients, similarly showed no effect of early blood pressure lowering upon death or dependency at 3 months, RR of 1.04 (95% CI: 0.96–1.13) (Lee et al., 2015).

Figure 5.1 Forest plot showing the effects of more intensive blood pressure lowering therapy in acute ischaemic stroke patients on death or dependency. Reproduced from Bath et al. (2014), with permission from John Wiley & Sons Limited.

Copyright Cochrane Library.

In most of the blood pressure lowering trials, treatment started in the late acute period, a median of 11 to 58 hours after stroke onset. The effects of blood pressure lowering in the early acute period, the first 0–12 hours after stroke onset, have not been as well studied in randomized trials and may differ from later timepoints, since substantial ischaemic tissue at risk (penumbra) is often still present and could be jeopardized by collateral blood flow reduction with blood pressure lowering. In the RIGHT-2 trial of a mixed blood pressure lowering and neuroprotective agent (glyceryl trinitrate) begun hyperacutely in the prehospital setting (median 70 min after onset), the common odds ratio point estimates for shift in 3-month disability outcome were non-significantly unfavourable for both ischaemic stroke (580 patients, common odds ratio [cOR]: 1.15, 95% CI: 0.85–1.54) and transient ischaemic attack (105 patients, cOR: 1.57, 95% CI: 0.74–3.35) (RIGHT-2 Investigators, 2019).

Gentle blood pressure lowering to systolic blood pressure (SBP) less than 185 and diastolic blood pressure (DBP) less than 110 in acute ischaemic stroke patients eligible for intravenous (IV) tissue plasminogen activator (tPA) therapy was required in the IV tPA treatment trials, and is currently recommended as a concomitant therapy in lytic-treated patients in national guidelines. However, the optimal blood pressure target for acute ischaemic stroke patients treated with IV tPA has not been conclusively defined by controlled clinical trials.

Comment

Avoiding extremes of blood pressure, including frank hypotension (SBP < 90 mm Hg) and hypertension sufficient to induce hypertensive encephalopathy (SBP > 220 mm Hg), seems prudent in acute ischaemic stroke patients to avoid deleterious effects on the brain and other organs. So far, there is no convincing evidence that blood pressure adjustments within this broad range during acute ischaemic stroke are beneficial.

A reasonable general strategy for patients with persisting vessel occlusions is to avoid lowering the blood pressure during the first 24 hours after stroke onset, unless a strong indication is present. During the first few hours after stroke onset, collateral blood flow to penumbral brain regions is most critical, and randomized trials have not specifically tested therapies in the hyperacute period. Thereafter, gradual blood pressure lowering may be safely started, and long-term, secondary prevention blood pressure targets eventually attained. There is little evidence to recommend a specific blood pressure lowering drug class during acute stroke. However, drugs that can control the blood pressure with minimal fluctuations may be preferred.

In patients in whom IV thrombolysis is initiated, stricter control of blood pressure during the first 24 hours is recommended to reduce risk of haemorrhagic transformation in accord with regimens used in lytic trials. In patients in whom partial reperfusion has been achieved with IV thrombolysis or endovascular thrombectomy, an individual judgement needs to be made about the best blood pressure target that balances the countervailing physiological goals of lowering blood pressure to reduce risk of reperfusion haemorrhage versus maintaining an elevated blood pressure to support residual hypoperfused fields.

Induced hypertension to enhance collateral flow in select patients with persisting vessel occlusions is a strategy that has shown promise in small trials, but has not been confirmed in definitive larger studies. Blood pressure augmentation might be used cautiously when there is evidence of persisting penumbral tissue in jeopardy of infarction or fluctuating stroke deficits in patients without congestive heart failure or coronary artery disease.

Dehydration

Both dehydration and overhydration can have detrimental effects during acute ischaemic stroke. Dehydration can lead to hypotension and subsequent hypoperfusion of vulnerable ischaemic brain regions, thus worsening a stroke. Overhydration can lead to congestive heart failure, especially in the elderly and those with renal or cardiac insufficiencies. In addition, volume overload may exacerbate ischaemic brain oedema.

Hydration status can be assessed by monitoring the mucous membranes, skin turgor, blood haematocrit, blood urea nitrogen/creatinine ratio, serum sodium, urine osmolarity, and body weight. For some critically ill stroke patients, more definitive measurements, such as central venous pressure or pulmonary capillary wedge pressure, are needed.

Evidence

Dehydration is common in acute ischaemic stroke patients. In a study of 2158 consecutive acute ischaemic stroke patients, 61% were dehydrated on presentation or at some point during their hospital course, as assessed by a blood urea nitrogen/creatinine ratio greater than 80 (Rowat et al., 2012). Patients with dehydration substantially more often died in hospital or were discharged to institutional care.

Two randomized trials have evaluated liquid support regimens in acute ischaemic stroke. In one trial, a broad population of 120 acute ischaemic stroke patients presenting within 72 hours of acute ischaemic stroke onset were randomized to a 0.9% NaCl solution at 100 mL/h for 3 days or to no IV liquids (Suwanela et al., 2017). There was no difference in the rates of freedom from disability (mRS 0–1) at 3 months with IV liquids versus control (83.3% vs 80.0%, p = 0.64). However, early neurological deterioration (NIHSS worsening by 3 or more points during the first 3 days), not of metabolic or haemorrhagic origin, occurred less often in the patients receiving IV liquids (3% vs 15%, p = 0.02).

In another trial, 212 diabetic patients presenting with acute ischaemic stroke and chronic hyperglycaemia (HBA1c ≥ 7.0) were randomized to moderate versus minimal IV hydration with normal saline (Lin et al., 2018). The moderate hydration group received 300–500 mL bolus followed by 40–80 mL/h for 72 hours, while the minimal hydration group received 40–60 mL/h for the first 24 hours and 0–60 mL/h during 25 to 72 hours. Long-term outcomes were not assessed. However, early neurological deterioration (total NIHSS worsening by ≥2 points, consciousness or motor score worsening by ≥1 point, or any new neurological deficit) was less frequent with moderate than minimal hydration, 10.5% versus 33.6%, OR 0.21 (95% CI: 0.10–0.44, p < 0.05).

Comment

Background

Blood glucose can fluctuate extensively during acute stroke, especially in patients with diabetes mellitus. Both extreme hypoglycaemia and hyperglycaemia can cause neurological dysfunction and seizures. Hyperglycaemia is considerably more common and persistent than hypoglycaemia during acute stroke. Approximately 40% of acute ischaemic stroke patients have hyperglycaemia above 130 mg/dL on admission (Williams et al., 2002). There is considerable evidence linking hyperglycaemia during acute stroke with worse functional outcomes (Capes et al., 2001; McCormick et al., 2008). However, hyperglycaemia during acute ischaemic stroke may be only a marker of more severe illness and underlying insulin resistance. Clinical trials to improve functional outcomes by reducing hyperglycaemia during acute ischaemic stroke have not demonstrated efficacy.

Evidence

A systematic review identified 11 randomized controlled trials enrolling 1583 patients to greater versus lesser control of hyperglycaemia during acute ischaemic stroke using different glucose-monitoring frequencies and insulin regimens (Bellolio et al., 2014). There was no difference between the intervention and control groups in the death or dependency outcome, OR 0.99 (95% CI: 0.79 to 1.23) (Figure 5.2). The rate of symptomatic hypoglycaemia was higher in the intervention group, OR 14.6 (95% CI: 6.6–32.2). Similar findings were observed in patient subgroups with and without diabetes mellitus.

Figure 5.2 Forest plot showing the effects of insulin therapy for glycaemic control compared with control treatment in acute ischaemic stroke patients on death or dependency. Reproduced from Bellolio et al. (2014), with permission from the authors and John Wiley & Sons Limited.

Copyright Cochrane Library.

More than half of the patients in the meta-analysis were enrolled in the Glucose Insulin Stroke Trial – United Kingdom (GIST-UK), in which 933 acute ischaemic stroke patients with mild-to-moderate hyperglycaemia were randomly assigned to treatment with IV insulin or saline (control) (Gray et al., 2007). There was no evidence of efficacy from the IV insulin treatment. However, the mean blood glucose was only 10 mg/dL lower in the intervention group than in the control group, and 84% of the subjects did not have diabetes mellitus, suggesting that undertreatment and enrolment of patients with stress hyperglycaemia rather than underlying diabetes mellitus or prediabetes may have masked a potential treatment benefit.

The Stroke Hyperglycemia Insulin Network Effort (SHINE) trial (Bruno et al., 2014), enrolled 1151 hyperglycaemic acute stroke patients, randomized to standard subcutaneous insulin (target blood glucose <80–179 mg/dL) or intensive IV insulin (target blood glucose 80–130 mg/dL) therapy for the initial 72 hours after stroke onset. Eighty percent of patients had a history of diabetes mellitus type 2. The mean blood glucose level in the standard treatment group was 179 mg/dL and in the intensive group it was 118 mg/dL during the treatment period. The rates of favourable functional outcome at 3 months were similar in the two treatment groups, 21.6% in standard and 20.5% in intensive, RR 0.97 (95% CI: 0.87–1.08) (Johnston et al., 2019). Occasional severe hypoglycaemia (<40 mg/dL) occurred only in the intensive group, in 2.6% of patients.

Comment

Extremely low and high blood glucose deviations should be avoided in patients with acute ischaemic stroke. Peripheral nervous manifestations of hypoglycaemia (tremulousness, palpitation, diaphoresis, paraesthesia) usually begin to appear when the blood glucose drops below 70 mg/dL (Cryer et al., 2003). Central nervous manifestations of hypoglycaemia (generalized weakness, fatigue, confusion, and eventually seizures) usually begin to appear when the blood glucose drops below 55 mg/dL (Cryer et al., 2003). The blood glucose threshold for developing manifestations of hypoglycaemia varies between patients, and those with diabetes tend to develop manifestations at higher glucose levels than those without diabetes. Treatment of mild-to-moderate hypoglycaemia could consist of juice consumption if not contraindicated by dysphagia. Otherwise, and to treat severe hypoglycaemia, IV 50% dextrose should be administered urgently to prevent seizures or hypoglycaemic neurological injury.

There is no agreed-on definition of hyperglycaemia during acute illness. Patients with type 2 diabetes mellitus rarely develop diabetic ketoacidosis. However, blood glucose levels above 260 mg/dL have been reported to cause focal neurological deficits resembling stroke, some with non-convulsive focal seizures. Non-convulsive hyperglycaemic hemianopia (Stayman et al., 2013; Strowd et al., 2014), aphasia (Huang et al., 2014), and mild hemiparesis (Hansford et al., 2013) have been reported.

Tight control of hyperglycaemia with IV insulin to a blood glucose range of 80–110 mg/dL is challenging and carries a significant risk of dangerous hypoglycaemia. A less-stringent control of hyperglycaemia is more feasible and less risky.

Based upon current clinical data, initial treatment of hyperglycaemia during acute ischaemic stroke using a subcutaneous regular insulin sliding scale of moderate intensity is appropriate. With current knowledge, a target blood glucose of less than 200 mg/dL during acute ischaemic stroke seems reasonable. If subcutaneous insulin treatment is insufficient to control an individual patient’s hyperglycaemia, IV insulin may be used, accompanied by more intensive blood glucose and patient monitoring. As the daily insulin requirement becomes more consistent, long-acting (basal) insulin doses may be introduced to minimize blood glucose fluctuations. After establishing regular oral or tube feeding diets, oral hypoglycaemic agents may be used more safely. Monitoring the blood glucose 4 times daily should be sufficient for most acute stroke patients with diabetes.

Dysphagia

Evidence

Randomized trials of dysphagia screening prior to start of oral feeding in acute ischaemic stroke patients have not been conducted, but multiple observational studies are supportive. In an analysis of 15 hospitals with 2532 acute ischaemic stroke patients, the 6 hospitals that routinely used a formal dysphagia screen before oral feeding had significantly lower rates of in-hospital pneumonia (2.4% vs 4%, p = 0.002) (Hinchey et al., 2005).

In a before–after study of 384 acute stroke patients, introduction of a formal 24/7 dysphagia screening by bedside nurses compared with prior weekday-only screening by speech and language pathologists was associated with earlier determination of swallowing safety (median 7 vs 20 hours) and significantly lower pneumonia rates (3.8% vs 11.6%, p = 0.004) (Palli et al., 2017).

A cluster-randomized trial of a nursing quality improvement intervention targeting fever, glucose management, and dysphagia screening reduced death and dependency, but without reducing the pneumonia rate (Middleton et al., 2011).

Another small cluster-controlled trial evaluated 162 acute stroke patients cared for on 2 wards, one implementing a formal swallow screen protocol and the other continuing with physician clinical assessment of swallowing (Rai et al., 2016). Stroke patients who received the formal swallow evaluations tended to have less aspiration pneumonia than the conventional care patients (6.5% vs 15.3%, p = 0.06).

Behavioural therapies to treat dysphagia, including swallowing exercises, oral stimulation, and oral care led by speech and language pathologists or nurses, have been tested in several randomized trials, generally comparing more intensive to less intensive therapies or standard care (Bath et al., 2018). Patients predominantly had ischaemic stroke, but some had haemorrhagic strokes. Across 6 randomized trials enrolling 511 patients, more active behavioural therapies were associated with less persistent dysphagia, 34.5% versus 53.5%, OR 0.45 (95% CI: 0.28–0.74); and across 6 randomized trials enrolling 473 patients, more active behavioural therapies were associated with reduced pneumonia, 20.3% versus 34.2%, OR 0.56 (95% CI: 0.31–1.00) (Figure 5.3).

Figure 5.3 Forest plot showing the effects of various swallowing therapies in acute and subacute stroke patients on pneumonia rates. Reproduced from Bath et al. (2018), with permission from the authors and John Wiley & Sons Limited.

Copyright Cochrane Library.

Acupuncture to treat dysphagia after stroke has been evaluated in numerous small randomized trials. A systematic review identified 59 randomized trials enrolling 4809 patients, most ischaemic but some haemorrhagic (Ye et al., 2017). Acupuncture was associated with more frequent improvement in swallowing function compared with standard therapies, 92% versus 77%, RR 1.17 (95% CI: 1.1–1.21). However, many studies had risk of bias due to incomplete blinding of patients and evaluators.

Pharyngeal electrical stimulation therapy to treat dysphagia, including electrical stimulation applied via electrodes implanted in a nasogastric tube, has been evaluated in at least 6 randomized trials (Suntrup et al., 2015; Bath et al., 2018; Dziewas et al., 2018). In 4 trials enrolling 177 patients, pharyngeal electrical stimulation was associated with non-significantly improved scores on the penetration-aspiration scale (Bath et al., 2018). A more recent multicentre trial, enrolling 69 severely dysphagic patients with tracheostomy, was stopped early for efficacy when readiness for decannulation (removal of tracheostomy) 24–72 hours after treatment occurred more frequently in the pharyngeal stimulation group, 49% versus 9%, OR 7.00 (95% CI: 2.41–19.88; p = 0.0008) (Dziewas et al., 2018).

Transcranial direct current and magnetic stimulation therapies added to standard therapies have been evaluated as dysphagia treatments in several trials. In a study-level meta-analysis of 8 randomized trials of transcranial magnetic stimulation, enrolling 141 patients, stimulation improved scores on swallowing scales (Bath et al., 2018). Analysis of 2 randomized trials of transcranial direct current stimulation, enrolling 34 patients, showed a lesser, non-significant, but directionally favourable effect (Bath et al., 2018).

Nutrition

Under-nutrition is present in 8–25% of acute stroke patients during the first week and is associated with worse functional outcomes and higher mortality (FOOD Trial Collaboration, 2003; Yoo et al., 2008).

The largest clinical study evaluating nutritional support strategies in acute stroke patients was the Feed Or Ordinary Diet (FOOD) study, which comprised a family of 3 multicentre randomized trials. One trial included acute stroke patients without dysphagia (FOOD Trial Collaboration, 2005a), and two included patients with dysphagia (FOOD Trial Collaboration, 2005b).

In the nutrition trial in individuals without dysphagia, 4023 acute stroke patients were randomized to eat either a regular hospital diet or a hospital diet supplemented with protein (FOOD Trial Collaboration, 2005a). At 6 months, there was no significant difference in the rates of dependency or death (mRS scores 3–6) between the two treatment groups (58% vs 59%).

One trial in patients with dysphagia compared early (≤7 days from hospital admission) versus late (>7 days from hospital admission) start of enteral tube feeding. Among 859 randomized patients, earlier feeding was associated with a trend toward reduced mortality at 6 months (42% vs 48%, p = 0.09), but with no increase in the rate of being capable of bodily self-care (mRS scores 0–3) (21% vs 20%, p = 0.7) (FOOD Trial Collaboration, 2005b). The rate of gastrointestinal bleeding was higher in the early feeding group (5.1% vs 2.6%, p = 0.04).

The second trial in dysphagia patients compared initial nasogastric (NG) versus initial percutaneous endoscopic gastrostomy (PEG) tube feeding. Among 321 randomized patients, allocation to NG versus PEG tube feeding was associated with higher rates of being alive and capable of bodily self-care at 6 months (mRS scores 0–3) (89% vs 81%, p = 0.05). Also, gastrointestinal bleeding was less frequent in the NG tube group (3.1% vs 11.3%, p = 0.005).

Comment

Observational and limited controlled trial data support performing simple, formal swallow screening early in all ischaemic stroke patients, with initiation of oral intake in patients with retained swallowing ability, and referral to speech therapists for more detailed assessment of patients with evidence of potential dysphagia. Beginning oral feeding only in patients cleared by a careful swallowing assessment will minimize aspiration complications. Based on the findings by a specialist in swallow evaluation, some acute stroke patients with mildly impaired swallow may be cleared initially to eat only mechanically modified foods, such as puréed foods, before advancing to regular food.

When swallowing is impaired so that adequate oral feeding is unsafe, initiating feeding via NG tube is reasonable within the first 2 or 3 days after stroke onset. If swallowing function has not improved sufficiently for adequate oral intake after 1 or 2 weeks, converting to PEG tube feeding is reasonable, as NG tube irritation of nasal mucosa becomes more likely. Tube feeding during acute stroke can be continuous or by periodic bolus, depending partly on patient tolerance. If swallowing improves during stroke rehabilitation, removal of the PEG tube should be considered. Behavioural swallowing rehabilitation therapies likely hasten resolution of dysphagia. Electrical pharyngeal stimulation may be of benefit for patients with prolonged dysphagia and tracheotomy.

Background

Fever is commonly defined as core body temperature at or above 37.5°C (99.5°F) and is relatively common during acute stroke (Grau et al., 1999; Middleton et al., 2011). Fever may be a sign of infection, evidence of other medical complications, or due to the stroke itself by a variety of mechanisms. Large areas of tissue necrosis can elevate body temperature (Reith et al., 1996). Extensive preclinical data indicate that fever enhances ischaemic brain injury, in part by accelerating the biochemical reactions mediating molecular cellular damage. Accordingly, lowering the body temperature to normal might slow deleterious biochemical reactions during acute ischaemic stroke and improve outcome. Furthermore, induced hypothermia is a potential neuroprotective intervention in acute ischaemic stroke (see Chapter 12).

Evidence

Patients with fever during acute stroke have worse outcomes than do patients without fever (Reith et al., 1996; Castillo et al., 1998; Prasad and Krishnan, 2010). The relationships between fever and stroke outcome or fever and infarct volume are strongest within the initial 6- to 24-hour period (Castillo et al., 1998), while in the first 3 hours after stroke onset, lower body temperatures are actually associated with worse outcome, likely from sympathetic hyperactivity initially lowering the hypothalamic thermoregulatory set-point (Kim and Saver, 2015). Fever may be a marker of stroke severity or a contributor to brain injury, or both.

A systematic review identified 7 randomized trials testing prophylactic versus reactive administration of pharmacological agents that reduce fever and maintain normothermia in acute stroke patients, predominantly patients with ischaemic stroke (den Hertog et al., 2009). Among 280 randomized patients, 61% were enrolled in 5 trials testing acetaminophen (paracetamol), 24% in 1 trial testing metamizol, and 16% in 1 trial testing ibuprofen. Overall, no differences were observed in the rate of dependency or death (mRS 3–6) between prophylactic and reactive pharmacotherapy, 50% versus 51%, OR 0.89 (95% CI: 0.54–1.48). In a larger, more recent trial, among 256 acute stroke patients randomized to high-dose paracetamol (6 g/day) or placebo for 3 days, body temperature from baseline to 24 hours in the control group increased by 0.18°C and in the intervention group decreased by 0.09°C (de Ridder et al., 2017). However, no benefit in global disability across mRS levels at 3 months was observed, with a common odds ratio of 1.02 (95% CI: 0.66–1.58).

While prophylactic antipyretic therapy has not been supported, a cluster-randomized trial provided support for close monitoring for fever development and aggressive reactive therapy for temperature elevation. In a cluster-controlled trial of a combined nursing quality improvement intervention targeting fever, glucose management, and dysphagia screening, the temperature intervention consisted of monitoring body temperature every 4 hours for the first 72 hours after admission and starting paracetamol if temperature was above 37.5°C (Middleton et al., 2011). A total of 19 acute stroke units were randomized, and 1696 patients enrolled. Patients in the intervention units were less likely to be dead or disabled (mRS scores 2–6) at 90 days (42% vs 58%, p = 0.002).

Comment

It is essential to monitor patients with acute ischaemic stroke for fever development. In patients developing hyperthermia, antipyretic therapy should be started immediately, whatever the cause. In addition, the underlying cause of the temperature elevation should be urgently investigated and directly treated. Usual treatments of fever include antipyretic medications and surface cooling devices (cooling blankets and pads). In addition, endovascular cooling devices are available. Further studies are needed to determine or clarify the degree to which fever is a contributor to poor outcomes versus a marker of more severe illness and to define optimal temperature targets.