Tableaux de données probantes et liste de référence
Medical complications are relatively common following stroke and are associated with increased lengths of stay and higher cost. Appropriate investigations and management strategies should be implemented for all hospitalized patients to avoid complications, prevent stroke recurrence and improve the odds of a good recovery. Estimates of the percentage of patients who experience at least one medical complication during hospitalization vary widely from 25% (Ingeman et al. 2011) to 85% (Langhorne et al. 2000). Some of the most commonly-cited complications include urinary tract infections, fever, pneumonia, and deep vein thrombosis (Otite et al. 2017, Indredavik et al. 2008, Roth et al. 2001).
Detecting atrial fibrillation (AF) after a stroke or TIA is important since it is a major risk factor for subsequent stroke and, once identified, can be effectively treated. However, AF is under-diagnosed because it is frequently paroxysmal and asymptomatic. Additionally, although many abnormalities can be detected within the first few days of monitoring, prolonged screening may be required to detect others. Flint et al. (2012) followed 239 patients with cryptogenic ischemic stroke who underwent outpatient cardiac monitoring using an electrocardiographic loop recorder for 30 days. Paroxysmal atrial fibrillation (PAF) was detected in 26 patients (11.0%; 95% CI : 7.6% to 15.7%) who were previously undiagnosed. While PAF was detected most often (45%) in patients within the first 10 days, 31% were detected from day 11 to 20 and 24%, from day 21 to 30. Suissa et al. (2012) included 946 patients with acute ischemic stroke who were previously undiagnosed with AF. Patients were admitted to an intensive stroke unit care that included continuous cardiac monitoring or to a conventional stroke unit care where patients received a baseline ECG, 24-hour Holter monitor and additional ECGs when necessary. Significantly more cases of AF were detected in patients in the continuous cardiac monitoring group (14.9% vs. 2.3%, adj OR=5.29; 95% CI 2.43 to 11.55). The odds of detection were highest within the first 24 hours of monitoring (OR=9.82; 95% CI 3.01 to 32.07). A prospective cohort study that compared the effectiveness of serial ECGs and Holter monitoring for the identification of AF in patients post stroke found that both methods were equally effective in identifying cases that were not present on a baseline assessment (Douen et al. 2008). Together, serial ECG’s and Holter monitoring identified 18 new cases of AF after baseline ECG assessment in the 144 patients included in the study. The majority (83%) of these cases were identified within 72 hours. A recent systematic review (Kishore et al. 2014) includes the results from 32 studies (5,038 patients) of patients with acute ischemic stroke or TIA who had undergone invasive or non-invasive cardiac monitoring for a minimum of 12 hours following event. The different types of cardiac monitoring evaluated included inpatient cardiac monitoring, 24, 48 & 72hr and 7-day Holter, external loop recorder, invasive cardiac monitoring and mobile cardiac outpatient telemonitoring. The overall detection rate of AF was 11.5% (95% CI 8.9%-14.3%) and was higher in selected (pre-screened or cryptogenic) patients (13.4%, 95% CI 9.0%-18.4%) compared with unselected patients (6.2%, 95% CI 4.4%-8.3%). The detection rate of AF in cryptogenic stroke was 15.9% (95% CI 10.9%-21.6%).
The use of transesophageal echocardiography (TEE) has been shown to be more sensitive compared with transthoracic echocardiography (TTE) for detecting cardiac abnormalities following ischemic stroke or TIA, although it is costlier and less acceptable to patients. Common TEE findings following stroke have included atheromatosis, patent foramen ovale, atrial septal aneurysm, (Marino et al. 2016, Katsanos et al. 2015). Marino et al. (2016) reported that 42.6% of 263 patients admitted following an acute ischemic stroke had a TEE finding which could explain the etiology of stroke/TIA. De Bruijn et al. (2006) included 231 patients with recent stroke (all types) or TIA whose stroke etiology remained in questions following initial ECG, ultrasound assessments and blood tests. All patients had a TEE followed by a TTE and the identification of major and minor cardiac sources of embolism were compared between the two diagnostic tools. A potential cardiac source of embolism was detected in 55% of the patients. Significantly more abnormalities were identified using TEE. A cardiac source was detected in 39% of patients where TEE was positive and the TTE, negative. A major cardiac risk factor was detected based on TEE in 16% of patients. The detection of possible cardiac sources of embolism was statistically significantly greater using TEE compared to TTE in both patients aged ≤45 years and >45 years.
Venous Thromboembolism Prophylaxis
The use of low molecular weight heparins (LMWH) has been shown to be more effective for the prevention of venous thromboembolism compared with unfractionated heparin (UFH) and is associated with a lower risk of serious bleeding events. A Cochrane review (Sandercock et al. 2008) included the results from 9 RCTs (n= 3,137) of patients with acute ischemic stroke who were randomized within 14 days of stroke onset to receive LMWHs or heparinoids, or UFH for an average of 10 to 12 days. The odds of DVT occurrence during treatment period were lower in the LMWH/heparinoid group (OR=0.55, 95% CI 0.44 -0.70, p<0.0001). There was no difference between groups in mortality during the treatment period or follow-up, nor in the odds of any ICH/hemorrhagic transformation during treatment (OR= 0.75, 95% CI 0.46- 1.23, p=0.25); however, there was an increased risk of major extracranial hemorrhage associated with the UHF group (OR= 3.79, 95% CI 1.30-11.06, p=0.015). The authors cautioned that the event rates for serious events (pulmonary embolus, death and serious bleeding) were too low to provide reliable estimates of the risk and benefits.
In the PREVAIL trial (Sherman et al. 2007), 1,762 patients who had experienced an ischemic stroke within the previous 48 hours and who were immobile with NIHSS (leg) motor scores of ≥2, were randomized to receive 40 mg enoxaparin subcutaneously once daily or 5000U UFH twice daily with UFH, for 10 days. The risk of all DVT at 14 days was 43% lower among patients receiving enoxaparin (10% vs. 18%, RR= 0.57, 95% CI 0.44 to 0.76, p<0.0001). The incidences of all proximal and distal DVT at 14 days were lower among patients receiving enoxaparin (5% vs. 10%, RR= 0.47, 95% CI 0.31 to 0.72, p=0.0003 and 7% vs. 13%, RR= 0.52, 95% CI 0.37 to 0.74, p=0.0002, respectively). There were no differences between groups in the incidence of symptomatic DVT or PE at 14 days (DVT: <1% vs. 1%, RR=0.29, 95% CI 0.06-1.38, p=0.096; PE : <1% vs. 1%, RR= 0.29, 95% CI 0.02-1.39, p=0.059). The protective effects were maintained at day 30, 60 and 90, following treatment. There were no significant differences between groups in any of the bleeding outcomes: total bleeding events, symptomatic ICH, major extracranial hemorrhage, all-cause mortality at days 14 or 90. In subgroup analysis treatment was effective regardless of time to initiation of prophylaxis, diabetes, obesity, previous stroke, stroke severity (NIHSS score ≥ 14 vs. < 14), gender or age. Using data from the PREVAIL trial, Pineo et al. (2011) conducted an economic analysis associated with enoxaparin or UFH use in a hypothetical cohort of 10,000 acutely ill medical inpatients. Although the drug cost was higher ($260 vs. $59), enoxaparin was associated with an overall average net savings of $1096 per patient. The cost savings was highest for patients with more severe strokes (NIHSS score≥14). The increased cost of enoxaparin was off-set by the avoidance of additional medical costs associated with reduced event rates of DVT and PE.
Anticoagulants and antithrombotics should be avoided in the early period following intracerebral hemorrhage to reduce the risk of worsening the initial hematoma. Evidence related to the benefit of venous thromboembolism prophylaxis is not as strong for this subgroup of patients. Orken et al. (2009) randomized 75 patients with primary ICH to LMWH (Enoxaparin 40mg/d) or long compression stockings (CS) after the first 48 hours of symptom onset. Hematoma volumes were calculated on the initial and follow-up CTs with the ABC/2 method. There was no evidence of hematoma enlargement at 72 hours, 7 or 21 days in either group. In addition, no other systemic bleeding complications were observed in the LMWH group. Four asymptomatic DVTs were detected (3 in LMWH and 1 in CS group). Investigators calculated the rate of asymptomatic DVT and PE in ICH patients, at 4% and 2.5% in the LMWH group. Tetri et al. (2008) reviewed the charts of 407 patients admitted for ICH patients, of whom 232 had received anticoagulant therapy for DVT prophylaxis using enoxaparin. Three-month mortality was similar between groups-19% in the treated group compared to 21% in the group who did not receive prophylaxis. Hematoma enlargements occurred in 9% and 7% of the treated and untreated patients, whereas symptomatic venous thromboembolic complications were observed in 3% and 2% of patients, respectively.
The use of external compression stockings/devices have been investigated in a series of three large, related RCTs, the Clots in Legs Or sTockings after Stroke (CLOTS) trials. In CLOTS 1 (Dennis et al. 2009), 2,518 patients, admitted to hospital within 1 week of acute ischemic stroke or ICH and who were immobile were randomized to either routine care plus thigh-length graded compression stockings (GCS) or to routine care plus avoidance of GCS. Patients wore the garments day and night until they became mobile, were discharged, or there were concerns with skin breakdown. At 30 days there was no significant difference between groups in the incidence of proximal DVT (GCS 10.0% vs. avoid GCS 10.5 %). GCS use was associated with a non-significant absolute reduction in risk of 0.5% (95% CI -1.9% to 2.9%). The incidence of any DVT or PE was non-significantly lower in the GCS group (17.0% vs. 18.4%, OR=0.91, 95% CI 0.74-1.11), but the frequency of skin ulcers or breakdown were significant higher in the GCS group (5.1% vs. 1.3%, OR=4.18, 95% CI 2.40-7.27). The inclusion criteria for the CLOTS 2 trial (The CLOTS Trials Collaboration 2010) were similar to those of CLOTS 1. In this trial, 3,114 patients were randomized to wear thigh-length stockings or below-knee stockings while they were in the hospital, in addition to routine care, which could have included early mobilization, hydration, and/or the use of anticoagulants/antiplatelets. At 30 days, there was a significant reduction in the incidence of proximal DVT associated with thigh-length GCS (6.3% vs. 8.8%, adj OR=0.69, 95% CI 0.53-0.91, p=0.008). The incidence of asymptomatic DVT were also lower in the thigh length GCS group (3.2% vs. 4.8%, adj OR=0.64, 95% CI 0.44-0.93, p=0.02). The use of thigh-length GCS was associated with an increased risk of skin breakdown (9.0% vs. 6.9%, OR=1.33, 95% CI 1.031.73, p=0.03). Finally, in CLOTS 3 (Dennis et al. 2013) 2,876 patients were randomized to wear thigh length intermittent pneumatic compression (IPC) device or to no IPC at all times except for washing and therapy, for a minimum of 30 days. The mean duration of IPC use was 12.5 days and 100% adherence to treatment was achieved in only 31% in the IPC group. The incidence of proximal DVT within 30 days was significantly lower for patients in the IPC group (8.5% vs. 12.1%, OR=0.65, 95% CI 0.51-0.84, p=0.001, ARR=3.6 %, 95% CI 1.4%-5.8%). There were no significant differences between groups for the outcomes of: death at 30 days (10.8% vs. 13.1%, p=0.057), symptomatic proximal DVT (2.7% vs. 3.4%, p=0.269), or PE (2.0% vs. 2.4%, p=0.453). The incidence of any DVT (symptomatic, asymptomatic, proximal or calf) was significantly lower for IPC group (16.2% vs. 21.1%, OR=0.72, 95% CI 0.60-0.87, p=0.001). Skin breakdown was more common in IPC group (3.1% vs. 1.4%, OR=2.23, 95% CI 1.31-3.81, p=0.002). At 6 months, the incidence of any DVT remained significantly lower in the IPC group (16.7% vs. 21.7%, OR=0.72, 95% CI 0.60-0.87, p=0.001). The incidence of any DVT, death or PE also remained significantly lower for IPC group (36.6% vs. 43.5%, OR=0.74, 95% CI 0.63-0.86, p<0.0001).
Elevated body temperature in the early post-stroke period has been associated with worse clinical outcomes. A meta-analysis conducted by Prasad & Krishnan (2010), including the results from six studies demonstrated that fever within the first 24 hours of ischemic stroke onset was associated with twice the risk of short-term mortality (OR= 2.20, 95% CI 1.59–3.03). Fever may result from a secondary infection, such as pneumonia, or may have occurred as a cause of stroke (e.g. infective endocarditis). While interventions to reduce temperature may improve the viability of brain tissue and/or prevent other medical complications post stroke, efforts to reduce fever, through a wide range of modalities, including pharmacological agents, (paracetamol) and physical interventions (cooling blankets and helmets and endovascular treatments) have not been convincingly shown to be effective in reducing/avoiding poorer outcomes.
Frank et al. (2013) conducted a retrospective study of 6,015 ischemic stroke patients who were registered in Virtual International Stroke Trials Archive (VISTA). Patients who received paracetamol for the management of pain (n=1626) or fever (n=809) were compared to those who had not received the medication. In patients treated with paracetamol for fever or pain, there was no difference in the distribution of mRS scores at 90 days, the primary outcome, compared with patients who did not receive treatment, while the odds of pneumonia were significantly reduced (OR=0.73, 95% CI 0.56-0.94, p=0.017). However, among patients without pain or fever who were treated with paracetamol as a prophylactic measure, the odds of poor outcome were increased (mortality at 90 days: OR=1.59, 95% CI 1.13-2.23, p=0.008, mRS score 0-2 : OR=0.55, 95% CI 0.41-0.74, p<0.001 and recurrent stroke within 7 days: OR=3.57, 95% CI 1.37-9.32, p=0.009). The largest trial examining the use of pharmacological agents for the reduction of fever was Paracetamol (Acetaminophen) In Stroke (PAIS) trial (den Hertog et al. 2009). In this trial, 1,400 patients were randomized to receive 1 gram paracetamol, 6x daily for 3 days or placebo within 12 hours of symptom onset. While treatment with paracetamol did significantly lower body temperature by a mean of 0.26 °C, it was not associated with improvement beyond expectation (adjusted OR=1.20, 95% CI 0.96-1.50), the increased odds of a favourable outcome, or significant increases in QoL. Treatment with paracetamol was associated with a decrease in 14-day mortality (OR=0.60, 95% CI 0.36-0.90), but there was no difference at 3 months (OR=0.90, 95% CI 0.68-1.18). The PAIS 2 trial (De Ridder et al. 2017) was terminated after enrolling 26 of 1,500 planned patients. In this trial, high-dose (2 grams) or placebo was given for 3 days to patients with a temperature of ≥ 36.5o C. There was no significant difference between groups in the shift in mRS scores at 90 days associated with paracetamol (common adj OR=1.15, 95% CI 0.74-1.79). In a Cochrane review (den Hertog et al. 2009) included the results from 8 RCTs, 5 of which examined pharmacological agents (paracetamol, n=3, metamizole n=1, ibuprofen placebo n=1) versus placebo. Pharmacological treatment significantly reduced temperature at 24 hours following treatment (MD= -0.21, 95% CI -0.28, -0.15, p<0.0001), but was not associated with a reduction in the odds of death or dependency at 1-3 months (OR= 0.92, 95% CI 0.59- 1.42, p=0.69).In terms of physical methods to reduce fever, the feasibility of endovascular and surface cooling strategies was examined in the COOLAID trial (Oversen et al. 2013). In this trial, 31 patients admitted to an ICU in two hospitals with acute ischemic stroke were randomized to receive therapeutic hypothermia (TH) using endovascular or surface methods, or standard supportive care (n=14). Patients in the TH group had body temperature lowered to 33 degrees C and were maintained for 24 hours, while patients in the standard care group received acetaminophen if body temp exceeded 37.5 degrees C. There were significantly more episodes of bradycardia associated with the TH group, and a non-significant increase in the incidence of pneumonia (6 vs. 1, p=0.09), although there were no significant differences between groups in other cardiac adverse events or pulmonary adverse events, or death. The authors concluded that the treatment was feasible, but associated with serious complications, particularly in anesthetized patients receiving endovascular cooling. A Health Technology Assessment (Harris et al. 2012) examined the use of any form of non-invasive head cooling following TBI, and cardiac arrest. The most effective techniques for which there were adequate data (nasal coolant and liquid cooling helmets) indicated that intracranial temperature could be reduced by 1 °C in 1 hour.
Early mobilization post stroke is intended to reduce the risk of medical complications including deep vein thrombosis, pressure sores, painful shoulders, and respiratory infections. The potential benefits of early mobilization have been examined in several RCTs, with ambiguous results. One of the sources of variability among studies examining the issue, which may account for conflicting results, is differences in treatment contrasts. Early mobilization was defined as early as 12 hours following stroke to as long as 52 hours, while patients in the delayed group were mobilized from time periods ranging from 48 hours to 7 days. Small sample sizes (i.e. under- powered samples sizes) may also have contributed to null findings. In the Akerhus Early Mobilization in Stroke Study (AKEMIS) 65 patients were randomized to a very early mobilization (VEM) group or to a control group following ischemic or hemorrhagic stroke. Patients in both groups received standard stroke unit care. Patients in the VEM group were mobilized as soon as possible (within 24 hours post stroke), while patients in the control group were mobilized between 24 and 48 hours. The median time to first mobilization from stroke onset was significantly shorter for patients in the VEM group (13.1 vs. 33.3 hrs, p<0.001); however, there were no significant differences between groups on any of the outcomes of interest, including poor outcome at 3 months (mRS score of 3-6), death or dependency, dependency, or number of complications at 3 months. Diserens et al. (2011) randomized 50 patients with ischemic stroke to either an “early mobilization” group who were mobilized out of bed after 52 hour or to a “delayed mobilization” group where patients were mobilized after 7 days. While there were significantly fewer severe complications among patients in the early mobilization group (8% vs. 47%, p < 0.006), there were no significant differences between groups in the numbers of minor complications, neurological deficits, or blood flow modifications.
Several publications are associated with the A Very Early Rehabilitation Trial for Stroke (AVERT) trial. The safety and feasibility of an early mobilization intervention was established by Bernhardt et al. (2008) in Phase I. 71 patients were randomized to receive either very early and frequent mobilization (upright, out of bed, activity – 2x/day, for 6 days a week until discharge beginning within 24 hours of stroke), or usual multi-disciplinary stroke team care. There was a non-significant increase in the number of patient deaths in the early mobilization vs. delayed mobilization group at 3 months (21% vs. 9%, absolute risk difference = 12.0%, 95% CI, 4.3% to 28.2%, p=0.20). After adjusting for age, baseline NIHSS score and premorbid mRS score, the odds of experiencing a good outcome were significantly higher at 12 months for the VEM group (OR= 8.15, 95% CI 1.61-41.2, p<0.01), although not at 3 or 6 months. In AVERT II, examining medical complications associated with very early mobilization (VEM), Sorbello et al. (2009) reported there were no differences in the total number of complications between groups. Severe complications or stroke-related complications occurred in 91 patients in the control group compared with 87 in the VEM group. Cumming et al. (2011) reported that patients in the VEM group returned to walking significantly sooner than patients in the standard care group (median of 3.5 vs. 7.0 days, p=0.032). While there were no differences between groups in proportions of patients who were independent in ADL, or who experienced a good outcome at either 3 or 12 months, VEM group assignment was a significant, independent predictor of independence in ADL at 3 months and of good outcome at both 3 and 12 months. Pooling the results from both the AVERT and VERITAS trials, which used similar protocols for early mobilization, Craig et al. (2010) reported that, compared with patients receiving standard care, patients in the VEM group were more likely to be independent in activities of daily living at 3 months (OR= 4.41, 95% CI 1.36-14.32), and were less likely to experience immobility related complications (OR= 0.20, 95%CI 0.10-0.70). The most recent replication of AVERT examined the effectiveness of a protocol of more intensive, early out-of-bed activity. Bernhardt et al. (2015) randomized 2,104 adults (1:1) to receive early mobilization, a task-specific intervention focused on sitting, standing, and walking activity, initiated within 24 hours of stroke onset, or to usual care for 14 days (or until hospital discharge). The median time to first mobilization was significantly earlier in the early mobilization group (18.5 vs. 22.4 hrs, p<0.0001). Patients in the early mobilization group received significantly more out of bed sessions (median of 6.5 vs. 3, p<0.0001) and received more daily therapy (31 vs. 10 min, p<0.0001). However, significantly fewer patients in the early mobilization group had a favourable outcome, the primary outcome, defined as mRS 0-2, at 3 months (46% vs. 50%; adjusted OR=0.73, 95% CI 0.59-0.90, p=0.004). There were no significant differences between groups for any of the secondary outcomes (shift in distribution of mRS, time to achieve assisted- free walking over 50m, proportion of patients able to walk unassisted at 3 months, death or serious adverse events), nor were any interactions identified based on pre-specified sub groups for the primary outcome (age, stroke type, stroke severity, administration of t-PA, or geographical region of recruitment). Further analysis of AVERT data (Bernhardt et al. 2016), controlling for age and stroke severity, suggested that shorter, more frequent mobilization early after acute stroke was associated with improved odds of favorable outcome at 3 months, while increased amount (minutes per day) of mobilization reduced the odds of a good outcome.
Nutrition and Dysphagia
A standardized program for screening, diagnosis and treatment of dysphagia following acute stroke results has been shown to reduce the incidence of pneumonia and feeding tube dependency. Bedside screening may include components related to a patient’s level of consciousness, an evaluation of the patient’s oral motor function and oral sensation, as well as the presence of a cough. It may also include trials of fluid. Coughing during and up to one minute following test completion and/or “wet” or hoarse voice are suggestive of an abnormal swallow. Silent aspiration may occur in patients who do not cough or complain of any problems with swallowing or have no wet-sounding voice, highlighting the importance of dysphagia screen for all patients acutely following stroke.
Hinchey et al. (2005) evaluated adherence to screening for dysphagia and associated pneumonia among individuals with ischemic stroke in the United States and reported that pneumonia occurred less frequently among those who had received a dysphagia screen (2.4% vs. 5.4%). Similar results were found in a study by Lakshminarayan et al. (2010) in which unscreened patients were found to have a greater risk of developing pneumonia than patients who had passed a screen for dysphagia (OR=2.2; 95% CI 1.7-2.7). In contrast to these two studies suggesting that screening is associated with a lower incidence of pneumonia, Masrur et al. (2013) reviewed the records of 314,007 patients with ischemic stroke admitted to hospitals participating in the Get-with-the-Guideline Registry. The outcomes of patients who had received a standardized swallowing screen including bedside or instrumental methods, were compared with those of patients who had not been screened. 68.9% patients were screened for dysphagia, while 31.1% were not. Of the 5.7% of patients who developed post-stroke pneumonia within 48 hours of admission, patients who were screened for dysphagia were more likely to develop pneumonia compared with those who did not develop pneumonia (7.5% vs. 68.5%, p<0.001). This finding suggests that patients who were perceived to be at high risk of dysphagia/aspiration may have been screened preferentially compared with patients perceived to be at low risk. To wit, Joundi et al. (2017) reported that patients with mild strokes were less likely to be screened compared with those with moderate strokes (adj OR=0.51, 95% CI 0.41-0.64) using data from 6,677 patients included in the Canadian Stroke Registry.
Middleton et al. (2011), in a multi-centered cluster RCT including 19 large tertiary care facilities with acute stroke units, randomized 4,198 patients to receive care at institutions that had adopted nursing protocols to identify and manage 3 complications- hyperglycemia, fever and swallowing dysfunction or to a control facility. The dysphagia component included education and training in the use of the ASSIST screening tool. While the intervention was associated with a decreased frequency of death or dependency at 90 days (42% vs. 58%, p=0.002) and swallowing screening was performed more frequently in the intervention group (46% vs. 7%, p<0.0001), there was no difference between groups in the incidence of pneumonia (2% vs. 3%, p=0.82). Using UK registry data, Bray et al (2017) reported a higher risk of stroke-associated pneumonia (SAP) with increasing times to dysphagia screening and assessment. The overall incidence of SAP was 8.7% (13.8% for patients not screened, 8.0% for patients who were screened and 14.7% for patients who received a comprehensive assessment). Independent predictors of receiving a dysphagia screen have been reported to include older age, admission to specialized units, the presence of weakness, increased stroke severity, speech difficulties and treatment with thrombolysis (Joundi et al. 2017, Mansur et al. 2013).
The effectiveness of a variety of treatments for dysphagia management was recently the subject of a Cochrane review (Geeganage et al. 2012). The results from 33 RCTs examining acupuncture, behavioral interventions, drug therapy, neuromuscular electrical stimulation, pharyngeal electrical stimulation, physical stimulation, (thermal, tactile) transcranial direct current stimulation and transcranial magnetic stimulation, were included. Pooling of results was not possible for many of the outcomes due to small numbers of studies available evaluating similar interventions/outcomes. Death or dependency at end of trial was the primary outcome, although only two RCTs were included in the pooled result. The results were not significant (OR=1.05, 95% CI 0.63 to 1.75, p=0.86). Acupuncture and behavioural modifications were associated with reduction in the presence of dysphagia at the end of treatment. No significant treatment effect was associated with subgroup analysis by therapy type (behavioral interventions, drug therapy, and electrical stimulation) for the outcome of chest infections.
Dietary modifications, including altered textured solids and fluids and the use of restorative swallowing therapy, and compensatory techniques, are the most commonly used treatments for the management of dysphagia in patients who are still safe to continue oral intake. Unfortunately, there is little evidence direct evidence of their benefit. Carnaby et al. (2006) randomized 306 patients with dysphagia admitted to hospital within 7 days of acute stroke, to receive usual care, standard low-intensity intervention (composed of environmental modifications, safe swallowing advice and appropriate dietary modifications), or standard high-intensity intervention and dietary prescription (daily direct swallowing exercises, dietary modification), for up to one month. When the results from the high-intensity and low-intensity groups were combined and compared with the usual care group, patients in the active therapy group regained functional swallow sooner and had a lower risk of chest infections at 6 months.
Oral supplementation can be used for patients who are not able to consume sufficient energy and protein to maintain body weight, or for those with premorbid malnutrition. The FOOD trial (Dennis et al. 2005a) aimed to establish whether routine oral nutritional supplementation in patients who could safely swallow and were prescribed a regular hospital diet, was associated with improved outcome after stroke. 4,023 patients were randomized to receive or not receive an oral nutritional supplement (540 Kcals) in addition to a regular hospital diet, provided for the duration of their entire hospital stay. At 6-month follow-up, there were no significant differences between groups on the primary outcome of death or poor outcome (OR=1.03, 95% CI 0.91 to 1.17, p>0.05). The absolute risk of death or poor outcome was 0.7%, 95% CI -2.3 to 3.8. Only 314 (8%) patients were judged to be undernourished at baseline. The anticipated 4% absolute benefit for death or poor outcome from routine oral nutritional supplements was not evident. The FOOD trial results would be compatible with a 1% to 2% absolute benefit or harm from oral supplements. Results from RCTs examining nutrition-related outcomes suggest that oral supplements can increase the amount of energy and protein patients consume, and prevent unintentional weight loss (Gariballa et al. 1998, Ha et al. 2010).
For patients who cannot obtain nutrient and fluid needs orally, enteral nutrition may be required. The decision to use enteral support should be made within the first seven days post stroke. The largest trial that addresses both the issues of timing of initiation of enteral feeding and the choice of feeding tube was the FOOD trial (Dennis et al. 2005b), which included 1,210 patients admitted within 7 days of stroke from 47 hospitals in 11 countries. In one arm of the trial, patients were randomized to receive either a percutaneous endoscopic gastrostomy (PEG) or nasogastric (NG) feeding tube within 3 days of enrolment into the study. PEG feeding was associated with an absolute increase in risk of death of 1.0% (–10.0 to 11.9, p=0.9) and an increased risk of death or poor outcome of 7.8% (0.0 to 15.5, p=0.05) at 6 months. In the second part of the trial patients were randomized to receive feeds as early as possible or to avoid feeding for 7 days. Early tube feeding was associated with non-significant absolute reductions in the risk of death or poor outcome (1.2%, 95% CI -4.2 to 6.6, p=0.7) and death (15.8%, 95% CI -0.8 to 12.5, p=0.09) at 6 months.
The incidence of post-stroke seizure ranges from 5%-15%, depending on stroke etiology, severity, and location (Gilad, 2012). Hemorrhagic events and cortical lesions are associated with the highest risk of both first and recurrent seizure (Gilad et al. 2013). Evidence examining the effectiveness of pharmacological treatment for post- stroke seizures is limited. A recent Cochrane review (Sykes et al. 2014) sought studies including patients of any age recovering from ischemic stroke or ICH, suffering from any seizure type that evaluated antiepileptic drugs compared with a placebo or no drug for the primary and secondary prevention of post stroke seizures. Only a single trial (Gilad et al. 2011) was found. In this trial, 84 patients with spontaneous non-traumatic and non-aneurysmatic ICH were randomized to receive 800 mg/day valproic acid or placebo daily for one month, for primary seizure prophylaxis. At 1 year, there were 15 cases of new seizure. There were no differences in early (within 14 days of randomization) or late (>14 days) seizure between treatment groups (1 vs. 4, p=0.8 and 6 vs. 4, p=0.5, respectively). Van Tuijl et al. (2011) planned to recruit 200 patients with lobar ICH or ischemic stroke, with a cortical syndrome and mRS≥3 or NIHSS ≥6. Patients were to be randomized to receive either 1500 mg of levetiracetam daily or placebo, within 2 to 7 days following acute stroke for primary seizure prevention. Treatment was scheduled to continue for 12 weeks. The trial was stopped prematurely due to a failure to recruit sufficient numbers of patients. At the point the trial was stopped, only 16 patients had been recruited over a period of 16 months.
The use of antiepileptic medications for the secondary prevention of seizures has also been examined, although placebo-controlled trials are absent. Gilad et al (2007) randomized 64 elderly patients admitted to a neurological department after stroke who had experienced a first seizure to receive either lamotrigine (100 mg BID) or carbamazepine (300 mg BID). The number of patients who were seizure free at 12 months was non-significantly higher in the lamotrigine group (23 vs. 14, p=0.06). The total number of adverse events was significantly higher in the carbamazepine group (12 vs. 2, p=0.05), as was the number of withdrawals for adverse events (10 vs. 1, p=0.02).
To avoid the onset of urinary tract infections (UTIs), the use of indwelling catheters is largely discouraged in clinical settings and is typically limited to patients with incontinence that cannot be managed any other way. If used, the catheter should be changed or removed as soon as possible. Ersoz et al. (2007) reported that among 110 patients consecutively admitted for rehabilitation following stroke, 30 developed a symptomatic UTI during hospitalization. UTIs occurred more frequently in patients with indwelling catheters, compared with patients who could void spontaneously (7/14 vs. 23/96, p=0.041) and in patients with residual urine volumes of >50 mL (41.2% vs. 19.5%, p=0.039). Several infection prevention strategies that have been identified to prevent or delay the onset of catheter-associated UTIs include inserting the catheter using aseptic technique, correctly positioning the drainage tube and the collection bag, maintaining uncompromising closed drainage, achieving spontaneous voiding, and administering intermittent catheterizations.
The effectiveness of bladder-training programs, which typically include timed/prompted voiding, bathroom training, pelvic floor exercises, and/or drug therapy, has been evaluated in a small number of studies. In one, 42 patients admitted to a single acute stroke unit, were each patient was prescribed an individualized bladder program consisting of bladder scanning, intermittent catheterizations/ post-void residual regimen, non-invasive voiding strategies (e.g. pelvic muscle exercises) and/or drug therapy. The regimen was continued until the post-void urine residual was below 100 ml for three consecutive days (Chan et al. 2007). Eighty-four percent of all patients achieved urinary continence within the first month of stroke. Among this group, all females became continent, while 23% of the male patients did not. In a Cochrane review, Eustice et al. (2000) included the results of 9 RCTs (n= 674), examining the potential benefit of prompted voiding (vs. no prompted voiding) provided for 10 days-13 weeks. Prompted voiding was associated with a reduction in the number of incontinent episodes in 24 hours (MD= -0.92, 95% CI -1.32 to -0.53, p<0.0001). In another Cochrane review examining a broad range of treatments for urinary incontinence, including behavioral interventions, specialised professional input, complementary medicine, pharmacotherapy and physical therapy, Thomas et al (2008) reported that treatment was associated with a decreased risk of urinary incontinence (RR= 0.44, 95% CI 0.23-0.86, p=0.0017). The mean improvement in FIM bladder score of 35 women with stroke who were admitted to a rehabilitation unit following the implementation of a standardized bladder management program was significantly greater (2.8 vs. 1.6, p=0.01) than those who had been admitted prior to the initiation of the program (Cournan 2012). Thomas et al. (2014) conducted a cluster feasibility trial, Identifying Continence Options after Stroke (ICONS). Compared with usual care, the systematic voiding program was not associated with significantly increased odds of being continent at 6 or 12 weeks.
Physical weakness following stroke may prevent patients from independently completing their activities of daily living, including oral care. Poor oral care, combined with potential side effects of medication (e.g., dry mouth, oral ulcers, stomatitis), may contribute to a greater amount of bacteria in the mouth, leading to the development of pneumonia. Patients have also reported lower oral health-related quality of life as a result of poor or inadequate dental care following stroke (Schimmel et al. 2011). Therefore, on admission to hospital, all patients should have an oral/dental assessment to examine mastication, tooth wear, disease and use of appliances, following stroke.
However, few studies have examined interventions to improve oral hygiene in patients following a stroke. Kim et al. (2014) reported that patients admitted to a neurosurgical ICU and randomized to an intervention group that received daily oral hygiene had lower Plaque Index and Gingival Index scores, compared with patients in a control group. Lam et al. (2013) included 102 dentate patients admitted to a rehabilitation unit following ischemic stroke or ICH within the previous 7 days, with a Barthel index score of <70. Patients were randomized to receive oral hygiene instruction (OHI), + chlorhexidine (CHI) mouth rinse, or OHI + CHI + assisted tooth brushing, twice daily for 3 weeks. The mean plaque index and Gingival Bleeding Index scores of patients in the OHI+CHX and OHI+CHX+assisted brushing groups were improved significantly more than patients that only received instruction on oral hygiene. A Cochrane review conducted by Brady et al. (2006) included the results of 3 RCTs (n=470) that included patients receiving some form of assisted oral health care (OHC) within a healthcare facility. Treatments evaluated included oral health care plus timed tooth brushing, health care education and selective decontamination of digestive tract using an antimicrobial gel applied to the mucous membranes of the mouth several times per day. Due to the small number of studies and variability in treatments, pooled analyses were not possible. The use of decontamination gel was associated with a reduction in the incidence of pneumonia: (OR=0.20, CI 95% 0.05 to 0.84, p = 0.03). A single education session was not associated with a reduction in dental plaque tooth coverage, the presence of gingivitis, or denture-induced stomatitis at one or 6 months following training, but was associated with a significant reduction in denture plaque at both assessment points and higher knowledge scores among care providers.