Developments in treating the nonmotor symptoms of stroke

Abstract

Introduction:

Stroke is among the most common causes of disability worldwide. Nonmotor symptoms of stroke are common and disabling. Many are treatable, and intervention improves the quality of life for stroke survivors.

Areas covered:

Here the author summarizes evidence-based treatment of depression and other mood disorders, aphasia, hemispatial neglect, impairments of emotional communication and empathy, deficits in memory and other cognitive functions, sleep disorders, pain, fatigue, (1)and seizures resulting from stroke. The author focuses on treatments supported by randomized controlled trials (RCTs), from literature cited in Google Scholar, Embase, and Pubmed.

Expert opinion:

While behavioral rehabilitation is the most common intervention for many of the sequelae of stroke, relatively small RCTs support the use of noninvasive brain stimulation (transcranial direct current stimulation and transcranial direct current stimulation) and medications that facilitate neural plasticity and recovery. These noninvasive brain stimulation methods remain investigational for post-stroke symptoms. The strongest evidence for pharmacological intervention is in the domains of post-stroke mood disorders and epilepsy, but additional RCTs are needed to confirm the efficacy of selective serotonin reuptake inhibitors and other medications for improving recovery of cognition, language, and energy after stroke.

1. Introduction

Worldwide, more than 80 million people are alive today who have survived a stroke; one of four people over age 25 will experience a stroke over the course of their lives. Stroke-related deaths and disability result in the loss of over 116 million years of healthy life. (https://www.world-stroke.org/assets/downloads/WSO_Global_Stroke_Fact_Sheet.pdf) Although motor symptoms, such as hemiplegia, may be the most noticeable sequelae, nonmotor symptoms account for far greater loss of productivity and reduction in the quality of life.(2) The burden of these nonmotor symptoms is carried not only by stroke survivors, but also their caregivers. Many of these symptoms, such as seizures, depression, anxiety, aphasia, hemispatial neglect, and other cognitive deficits can be effectively treated, often with a combination of behavioral and pharmaceutical interventions with or without noninvasive brain stimulation. Noninvasive brain stimulation includes treatments such as Transcranial Magnetic Stimulation, in which a change in magnetic field causes electric current at a specific area of the brain, and Transcranial Direct Current Stimulation (tDCS), in which nonpainful, low amplitude (typically 1–4 milliamp) electrical stimulation over specific brain areas is used to reduce (or increase) the firing threshold of neurons to modulate neuroplasticity. These treatments remain investigations, except in depression, for which effectiveness has been adequately established through randomized clinical trials (RCT). Behavioral interventions range from counseling to computer-delivered rehabilitation programs, sometimes delivered remotely via internet. In this narrative review I discuss evidence-based interventions for the most prevalent, although not all, of the nonmotor symptoms of stroke, focusing on evidence from clinical trials. As dysphagia, or swallowing impairment, is largely a motor impairment, it will not be covered here, but the interested reader is referred to reference(3) for review of management of this important consequence of stroke. I will also not review the vast literature on vascular cognitive impairment, which can be due to subcortical white matter hypertensities rather than true stroke, but readers can refer to a recent review(4).

2. Mood

• 2.1.Depression is one of the most common consequences of stroke. It is also among the most effectively treated. Clinically significant depression occurs in about 23% (6–37%) percent of stroke survivors, depending on the community studied, the time post-stroke, and method of diagnosis.(5–8) Depression is generally diagnosed with depression scales, such as the Hamilton Depression Scale (8) or Patient Health Questionnaire-9 (9), or psychiatric interview.

  A number of medications have been shown to be effective in ameliorating depression, including selective serotonin reuptake inhibitors (SSRIs)(10–12), serotonin-norepinephrine reuptake inhibitor (SNRIs), tricylclic antidepressants, and those with mixed effects (6, 12–14). For example, an early randomized clinical trial (RCT) of nortriptyline showed significantly lower scores on the Hamilton Depression Scale (p=0.006) at 5 or 6 weeks after the start of treatment (13). Daily dosages were: 20 mg, for 1 week, 50 mg for 2 weeks, 70 mg for 1 week, and 100 mg for 2 weeks. Although effective, nortriptyline has many common adverse side effects, including dizziness, drowsiness, dry mouth, increased hunger and weight gain, as well as rare but serious side effects such as cardiac arrhythmia. SSRIs are generally better tolerated, but have common side effects of reduce libido, dizziness, anxiety, and diarrhea or constipation. One meta-analysis of 12 SSRI trials (with various doses and medications over 1–16 weeks) reported a weighted mean difference in the Hamilton Depression Scale of − 5.5 (− 8.3to-2.7) after treatment versus 0.5 (0.0–1.0) before treatment, with a rate difference of 0.23 (0.03–0.43) and a significant benefit after just 3–4 weeks of treatment(15). Another meta-analysis of seven SSRI trials and three tricyclic antidepressant trials reported that the standardized mean difference on the Hamilton Depression Scale between placebo and antidepressant at 6–26 weeks was − 0.53 (− 0.97 to − 0.09) for SSRIs and − 1.41 (− 2.51 to − 0.31) for tricyclic antidepressants.(16)

  Some of these antidepressants have been shown to have a positive effect on motor recovery ((17–19) but see(20)) or cognition(21), independently of the effects on depression. Other studies report improvement in depression with an SSRI, but no effects on more general function.(22) However, a meta-analysis indicated greater effects of early antidepressants on neurological outcome than depression.(23) Differential effects of antidepressants likely reflect differences in the sensitivity of outcome measures across studies as well as medications and their dosages. Other interventions for post-stroke depression that have been effective are transcranial direct current stimulation (tDCS (24, 25)), transcranial magnetic stimulation (TMS (26)), exercise(27, 28), and counseling(12). TMS and tDCS have also shown benefit in improving other functions after stroke, as described below, but may require different target areas of stimulation.

  It should be noted that post-stroke depression and, so-called “vascular depression” (29), are different entities compared to the early-onset recurrent major depressive disorder, in terms of clinical characteristics, severity, disease course, pathophysiology, and response to treatment, even though many of the treatments are the same.(30)

  Treatment of mood is critical to enable full participation in rehabilitation, maintain life style modifications essential to reduce risk of recurrent stroke, and re-entry into social, vocational, and avocational activities(31). A hopeful, energetic, and positive attitude toward recovery can have a major influence on both response to therapy and compliance with both medications and home practice.

• 2.2.Mania and bipolar disorders are less commonly reported after stroke, but can occur.(32, 33) The incidence and prevalence are unknown. One systematic review identified 49 studies, which together reported 79 cases.(34) They identified risk factors for development of mania post-stroke as male sex and right cerebral infarct. They typically had no personal or family history of psychiatric disorder, had at least one vascular risk factor, but no subcortical atrophy. Treatments include lithium, valproic acid, and lamotrigine(35), but there has been no RCT to confirm their effectiveness after stroke. Valproic acid and lamotrigine can have the added positive effects of reducing the likelihood of post-stroke seizures, and may have fewer adverse side effects in this population.
• 2.3.Anxiety is also common, especially in survivors with depression. In one study of post-stroke depression, and 39% of women and 26% of men had an associated anxiety disorder, primarily agoraphobia.(5) Anxiety can be effectively treated with SSRIs and/or counseling.(36) Natural light has been used to treat both anxiety and depression. For example, one quasi-randomized controlled trial of 90 stroke patients found positive effect of naturalistic lighting on rehabilitation unit compared to standard indoor lighting. Scores improved on both the Hamilton Depression Scale (p = 0.011) and anxiety rated on the Hospital Anxiety and Depression Scale (p = 0.045) with natural compared to standard indoor lighting.(37)
• 2.4.Apathy is also reported after stroke(38), particularly to the right hemisphere, but is difficult to identify in people with anosognosia, or impaired recognition of one’s deficits, another common deficit after right hemisphere stroke. (39) Therefore, its true incidence and prevalence are unknown. Right hemisphere stroke can cause a reduction in emotional response to both positive and negative events.(40) Treatment of apathy is notoriously difficult, and there are no medications that have a reliable effect. However, to the degree that apathy is a consequence of depression, antidepressant medications can be helpful.

3. Energy

• 3.1.Fatigue is arguably the single most common and long-lasting symptom of stroke (41–43), occurring in 29–77% of patients.(44) It is frequent even after other symptoms have recovered. In one study, the presence of fatigue was independent of depression, but depression influenced the degree to which fatigue impacted on functional abilities. (41)
• 3.2.Functional imaging studies have provided evidence for the neural mechanisms of post-stroke fatigue in people who seem to have recovered well. (45) When stroke survivors are able to carry out a task “normally”, much larger areas of the brain are activated compared to when healthy controls carry out the same task with the same proficiency. Recruitment of larger areas of the brain requires more energy. This change in activated areas likely reflects the fact that recovery demands that when there is a lesion in an areas normally required by the task, other areas must assume the function of the damaged ones. The areas recruited are not always as “efficient” in carrying out the task, because have not yet been frequently recruited into the neural networks that normally accomplish the task. It is likely that fatigue can be best combatted through a combination of frequent practice in the recovered functions (language, motor skills, etc.), combined with adequate sleep. More sleep might be required after stroke than prior to stroke, to replenish energy and consolidate new memories and skills.
• 3.3.The few randomized controlled trials of treatment to reduce post-stroke fatigue have been negative.(46) However, a pilot study of modafanil showed some promise in reducing post-stroke fatigue, and has been used to reduce fatigue in other neurological conditions. This RCT of modafinil 100 mg daily in 36 stroke survivors identified a significant decrease in fatigue measured with the multidimensional fatigue inventory in the modafinil group compared to placebo (−7.38; 95% CI, −21.76 to −2.99; P<0.001). (43) Amantadine, which has long been used to treat fatigue in multiple sclerosis, is also sometimes effective in treating post-stroke fatigue, although RCT in stroke are lacking. (47) Another promising approach is exercise. Based on the proposed mechanism of inefficient neural networks underlying task performance, practice of the desired tasks (activities of daily living, speaking, motor skills, etc.) should improve efficiency over time, and reduce or eliminate associated fatigue. It is critical to avoid the vicious cycle of fatigue begetting idleness, which further worsens fatigue. Given the importance of the problem, more trials are needed to establish effective interventions.

4. Language

Impairments of production and comprehension of spoken and/or written language (aphasia) are very common and disabling, predominantly after left hemisphere stroke.

Behavioral speech and language therapy is the mainstay of treatment, supported by a number of randomized controlled trials. (48) Nevertheless, recent clinical trials have shown that transcranial magnetic stimulation, transcranial direct current stimulation, and possibly medications can augment the effectiveness of speech and language intervention (see(49, 50) for review). The largest randomized, double-blind, sham-controlled trial of tDCS to augment language treatment for post-stroke aphasia targeted, with anodal tDCS, the area of greatest activation in the left hemisphere during spoken naming (identified by pre-treatment functional MRI). In this study of 74 individuals with chronic post-stroke aphasia, the tDCS and the sham groups had identical computer-delivered naming therapy for 15 sessions. TDCS was associated with greater change in number of correctly naming pictured objects, the primary outcome measure. There was a 70% greater improvement in correct naming for anodal tDCS relative to sham.(51) Response to tDCS is greatest in people with the normal met met allele of brain derived neurotrophic factor (BDNF) genotype, compared to val carriers(52). A small RCT of TMS for post-stroke aphasia used high frequency rTMS (stimulatory) applied to the injured left IFG and low frequency rTMS (inhibitory) to the right sided, homologous IFG or bihemispheric sham stimulation for a total of 10 sessions followed by 30 minutes of SLT. In this trial, participants who received rTMS showed significantly greater improvements compared to sham in accuracy of word comprehension (p=.04), naming (p=.01), repetition (p=.002), and in aphasia severity (1.8±1.2 vs 0.9±0.3; p=.018).(53)

A RCT of donepezil 10 mg compared to placebo in 26 patients with chronic post-stroke aphasia greater improvement on the primary outcome measure of the Aphasia Quotient on the Western Aphasia Battery in the donepezil group at the 16 week endpoint relative to the placebo group (p <0.037). However, the difference was small (62.6 vs 59.5 on a 100 point scale) (54). Similarly a RCT of memantine 10 mg compared to placebo twice daily in 28 patients with chronic post-stroke aphasia greater improvement on the primary outcome measure of the change in the Aphasia Quotient on the Western Aphasia Battery in the memantine group relative to the placebo group at each follow-up time point (week 16, p = 0.002; week 18, p = 0.0001; week 20, p = 0.005) and at the washout assessment (p = 0.041).(55) However, again the differences were small (<5 point difference in change on a 100 point scale). Thus, RCTs support the use of tDCS plus language therapy(51), donepezil, or memantine(55); but there have been too few trials of medications to recommend pharmacological interventions as standard practice. There have also been recent developments in behavioral therapies, sometimes incorporating remote (internet-based) therapies and/or computer-delivered therapies. (50)

5. Attention: hemispatial neglect

Neglect of the side of the body or space contralateral to stroke can be extraordinarily disabling, especially after right hemisphere stroke. Right neglect also occurs after left hemisphere stroke, and may be as common as left neglect after right hemisphere stroke, but is often more subtle and less disabling, as it tends to involve neglect of the right side of individual objects, stimuli, or words, irrespective of the side of the viewer.(56)

Treatments of neglect include behavioral therapies, computer programs, prism adaptation, tDCS, TMS, caloric or optokinetic stimulation, and ipsilesional neck vibration (see (39, 57) for review). Prism adaptation is the most well established behavioral approach to improving attention to the contralateral side of space, with several non-randomized clinical trials supporting its effects.(58–60) RCT are difficult, as there is no valuable “sham” for prism adaptation. The use of tDCS or TMS to alleviate neglect has also received substantial support(61), and can augment the effects of behavioral therapies, including prism adaptation.(62) For example, a meta-analysis of 12 RCTs (273 participants) and 4 non-RCTs (94 participants) showed a benefit in overall spatial neglect measured by the line bisection test with non-invasive brain stimulation in comparison to sham (standardized mean difference −2.35, 95% CI −3.72, −0.98; p = 0.0001).(63)

6. Emotional communication

Impairments in expression or recognition of emotion through variations in tone, rate, rhythm, and loudness of voice (prosody) are even more frequent (about 21%) than hemispatial neglect (about 18%) after right hemisphere stroke (64), although the prevalence varies widely, depending on how these deficits are identified and the age of the population. The spectrum of deficits in emotional communication is often underestimated and under-recognized. While some of these disturbances are specific to negative emotions (e.g. fear and disgust) after right temporal-insular stroke(65) or amygdala stroke(66), they usually affect recognition and/or expression of all emotions. There have been very few trials of treatment to improve prosody comprehension or production, but one trial showed promise of behavioral treatment to improve production of emotional prosody.(67) A recent study indicated that impaired emotional prosody expression and recognition (aprosodia) can be caused by impairment in knowing how an emotion (e.g. sad) “should” sound. Right hemisphere stroke patients with this deficit often improved by cues to re-establish this abstract knowledge (e.g. sad: low-pitched, quiet, slow, monotone).(68) These behavioral treatments require further investigation, and might be augmented by tDCS.

7. Empathy

Family members report that among the most disturbing of sequelae of right hemisphere stroke is the disruption of affective empathy – the recognition and sharing of the emotions of another.(69) The prevalence is unknown, because large epidemiological studies in stroke patients are lacking. Impaired emotional empathy occurs with damage to right cortical and limbic regions(70–72). Impairment in empathy is identified clinically though rating scales, such as the Interpersonal Reactivity Index (IRI), which is used as a self-rating scale in neurotypical controls, but can be given to spouses of people with neurological disease, as the patients may also have impaired insight(73). For research purposes, we and others have evaluated empathy using skin conductance response while watching others in positive or negative experiences(74). Although there have been no randomized trials of medications or other interventions to improve affective empathy after stroke, intranasal oxytocin is a theoretically-motivated intervention that is being tested in frontotemporal dementia.(75–77)

8. Memory and other cognitive functions

• 8.1.ICognitive dysfunction is common after stroke. It occurs as a direct result of the lesion (e.g. strokes involving the hippocampus can cause severe anterograde memory deficits), hypoperfusion around the lesion, and associated small vessel cerebrovascular disease that disrupts white matter connections between cortical regions.(78) Most functions that are negatively influenced by stroke improve in the months or years after stroke, but in about 25% of stroke survivors, cognitive function continues to decline. In these individuals, the stroke may have triggered vascular dementia or vascular cognitive impairment.(79) Although there is no disease modifying treatment of vascular cognitive impairment, one RCT showed a small but significant effect of donepezil in reducing the rate of cognitive decline after stroke.(80) Another RCT demonstrated a positive effect of escitalopram on cognitive function after stroke.(21) Cognitive rehabilitation provided by speech-language pathologists or occupational therapists can also be helpful, often in providing approaches to compensation. While treatment of vascular cognitive impairment and vascular dementia is beyond the scope of this paper, readers are referred to a recent consensus paper aimed at the developing disease-modifying strategies.(81)
• 8.2.Delirium is also common in the acute stage of stroke, particularly during hospitalization. In one study of 750 consecutive stroke patients (71.75 ± 13.13 years), 203 (27.07%) had delirium(82). Factors that contribute to delirium are disrupted sleep (e.g. for frequent neurological exams), polypharmacy, unfamiliar environment, and cognitive deficits. The most effective treatment is normalization of the environment, providing clear demarcations between day and night, simplification of medication regimens, and re-orientation. Associated hallucinations, delusions, and extreme agitation can be treated with atypical antipsychotics, such as olanzapine and quetiapine, but these dopamine receptor blockers (particularly the older ones such as haloperidol) should be avoided when possible, because they may interfere with neuroplasticity needed for recovery.(83) Dopamine is an essential neurotransmitter for long term plasticity and learning. Cannabidiol and medical cannabis may prove to be helpful in treating agitation, although RCTs are needed to evaluate the benefits and risks. These medications (which have different legal status across states in the US and across countries) can increase the anticoagulation effects of warfarin, and should not be used in patients on warfarin.

9. Seizures

About 10% of people with stroke experience one or more seizures. Seizures that occur at onset have low risk of recurrence, but seizures that begin later, usually around one year post-stroke, carry a high risk of recurrence.(84) Levetiracetam (500 to 1000 mg twice daily) is generally the drug of choice for secondary prevention of late-onset seizures, because it is effective and generally well tolerated, and has few interactions with other medications. A systematic review of RCTs of anti-epileptic medications revealed no significant difference in seizure freedom between levetiracetam and controlled-release carbemazepine (49/52 versus 46/54; p = 0.08) or between lamotragine and controlled-release carbemazepine (23/32 versus 14/32; p = 0.06). However, adverse effects occurred in fewer patients on levetiracetam than on controlled-release carbemazepine (17/52 versus 21/54; p = 0.02) and for those on lamotragine versus controlled-release carbemazepine (2/32 versus 12/32; p = 0.05). (85) Levetiracetam may even have other positive effects, such as neuroprotection or improvement in cognition, although the evidence for these effects is somewhat weak. Some individuals experience irritability with this medication, so other options include carbamazepine, lamotrigine, or lacosamide. Lamotrigine is particularly well tolerated, and may be as effective as levetiracetam.(86)

Although post-stroke seizures are caused by the focal lesion, the threshold for seizures is lowered by infections, poor sleep, and certain medications (e.g. many pain medications, donepezil, memantine). It is important for stroke survivors to obtain adequate sleep, rest, and treatment of infections.

10. Sleep disorders

Sleep disorders occur in up to 50% of patients after stroke and can negatively influence short‐term and long‐term functional outcomes, risk of stroke recurrence, and hospital length of stay.(87, 88)

• 10.1.Sleep‐disordered Breathing is the most common sleep disorder after stroke and includes obstructive sleep apnea (OSA) and central sleep apnea (CSA). The latter resembles Cheyne–Stokes breathing. Sleep-disordered breathing is characterized by night‐time symptoms of excess respiratory noises such as snoring, irregular breathing, sleep-onset insomnia, shortness of breath, palpitations, nocturia, agitated sleep and daytime symptoms of sleepiness, headaches, and impaired concentration and memory.(87) OSA can contribute to post-stroke fatigue. Risk factors for the development of OSA and CSA include central obesity, increased age, male sex, and neck circumference.(87) These disorders are diagnosed with polysomnography. CSA is most common after bilateral strokes, strokes associated with disturbed levels of consciousness, or strokes accompanied by heart failure.(41)

  Poor outcome associated with these disorders may be accounted for by elevated blood pressure levels, recurrent hypoxia, and cerebral hypoperfusion. Sleep disruption may also negatively impact neuroplasticity (which is upregulated during normal sleep).

  The mainstay of treatment of sleep apnea is Continuous Positive Airway Pressure (CPAP). A review of six RCT of CPAP to treat OSA revealed a significant improvement in daytime sleepiness and neurological function (measured with a variety of stroke scales).(87) However, compliance with CPA is often less than 50%; although some studies have achieved 70% compliance. Other interventions, such as dental appliances, to treat OSA are less effective but better tolerated.

• 10.2.Restless legs syndrome (RLS) has a reported increased prevalence after stroke, of about 12%, and occurs primarily after strokes involving the pons, thalamus, basal ganglia or corona radiata, and is infarcts. RLS symptoms are usually bilateral, but can be only contralateral to the stroke. The most common treatment of stroke‐related RLS is with dopaminergic agonists (e.g. ropinirole 0.25–1mg/day, pramipexole 0.125–0.5mg/day), but there have not been RCT.
• 10.3.Hypersomnia certainly occurs after stroke, but research is in this domain is quite limited, and the prevalence depends a great deal on the age the population and stroke severity. Although there have been no RCT, successful treatment reported in case studies and case series includes us of levodopa (100mg/day), methylphenidate (5–30mg/day), and modafinil (200 mg daily).
• 10.4Insomnia is common in all hospitalized patients, but may persist at home after stroke, largely due to anxiety and other mood disorders (see Section 1). Quiet environments, reduced procedures and exposure to light at night, increased light exposure and exercise during the day are practical and often useful approaches to management. If it is necessary to temporarily use hypnotic meducation, those without cognitive, antihistamine, or muscle‐relaxant effects are preferred, such as zolpidem, zopiclone, or sedating antidepressants, such as trazodone.(87) Benzodiazepines should be avoided, as they can impair recovery or even result in the re‐emergence of stroke symptoms that had previously recovered.

11. Pain

Pain is most common in the chronic stage of recovery from stroke, often due to spasticity, malaligned joints or shortened muscles. In one study of nearly 500 patients, the mean prevalence of pain after stroke at all stage was 30%, but varied by time point after stroke: 14% in the acute stage, 43% in the subacute stage, and 32% in the chronic stage, although different types of pain showed distinct time courses.(89) Post-stroke shoulder pain occurs in 50% to 80% of all individuals who have persistent upper extremity motor deficit.(90) Central pain syndromes occur in 1–12% of stroke survivors, and can be caused by altered sensitivity to somatic stimuli.(89) One type of central pain, thalamic pain syndrome, also known as Dejerine–Roussy syndrome, is a very severe type of neuropathic pain that occurs on the contralesional side after thalamic or internal capsule stroke.(91)

Treatment depends on the cause. Treatment of pain caused by spasticity can be treated by reducing spasticity (e.g. with baclofen 5–20 mg four times daily), tizanidine (2–6 mg up to four times daily), botulinum injections every three months, massage, strapping, slings, or other physical and occupational therapy interventions.(92) Common side effects of medications for spasticity include weakness, fatigue, headache,and nausea. Intrathecal baclofen may be more effective that oral antispasmodic medications.(93) Other interventions for post-stroke pain include minimization of glenohumeral subluxation, nerve blocks, and electrical stimulation.(92) Shoulder-hand syndrome is generally treated with physical and occupational therapy, although acupuncture has been used as an adjunct.(94) Treatment of Dejerine–Roussy syndrome is very difficult, but is sometimes ameliorated by sterotactic thalamotomy, deep brain stimulation or less completely with tricyclic antidepressants, gabapentin, pregabalin, or duloxetine. These last four medications are used to treat a variety of neuropathic pain symptoms. However, effectiveness and side effects increase with the dose. Common side effects of tricyclic antidepressants are discussed in Section 1.1. Common side effects of gabapentin and pregabalin include tingling, sedation, and lower extremity edema. Common side effects of duloxetine include light-headedness; visual changes, eye pain, swelling, or redness, and easy bruising.

The prevalence of headache, unlike other post-stroke pain syndromes, decreases over time. Headaches generally are due to hemorrhagic stroke acutely or those that result in increased intracranial pressure either acutely (e.g. due to edema) or chronically (due to hydrocephalus). Incidence of headaches in these settings approaches 100%. These headaches are often successfully treated with acetazolamide, although RCT are lacking. Chronic headaches should be treated in the same way as migraine headaches with a combination of prophylactic medications (antiepileptic medications, beta blockers, calcium channel blockers) and abortive treatments such as triptans.

10. Expert opinion

Treatment of nonmotor symptoms of stroke is essential to improve the quality of life of stroke survivors. While behavioral rehabilitation is the most common intervention for many of the sequelae of stroke, relatively small RCTs indicate that noninvasive brain stimulation (tDCS and TMS) and medications facilitate neural plasticity and recovery and augment behavioral interventions for many symptoms (see Table 1). Seizures, pain, and some sleep disorders are effectively ameliorated by medications, while OSA and CPA require other interventions, such as CPAP.

Although most trials of noninvasive brain stimulation for depression, aphasia, and neglect show a positive effect, the effect sizes are relatively small and long term benefits have not been adequately studied.(95, 96). tDCS is a relatively inexpensive modality of treatment with few risks or adverse effects, but there remain many questions about how to optimize the effectiveness. These questions include the optimum timing (e.g. at what time after stroke do patients benefit the most), the most effective site of stimulation (e.g. perilesional versus contralesional or cerebellar), the optimal amplitude of stimulation to maximize effects and minimize risk, the best number and frequency of treatment sessions, and the most effective approach (anodal perilesional, cathodal contralesional, or both simultaneously). Small studies have addressed some of these questions, but larger, multicenter comparative trials are needed to systematically study the influence of each of these variables. Furthermore, ways to maintain the effects of non-invasive brain stimulation are needed. Some investigators are studying monthly or bimonthly “maintenance” stimulation sessions. Others have considered the possibility of self-administered home tDCS with an appropriate task, to maintain gains achieved with clinician-administered tDCS with rehabilitation. However, concerns have been expressed by the international scientific community regarding the risks versus benefits of self-administered tDCS(97). Several studies have reported that gains in trained tasks are obtained with behavioral interventions alone (with sham tDCS), but that generalization to untrained tasks are observed only with simultaneous tDCS. These results provide further evidence that tDCS may have its effects through neuroplasticity within targeted functional networks. However, the extent of generalization to untrained tasks needs to be further explored. That is, for example, can the same intervention (tDCS to a particular site, with a given behavioral task) be used to improve motor functions, language functions, and mood? Or are effects specific to those tasks that depend on a particular neural network (e.g. the language network) or to those tasks that depend on the functions of the stimulated area (e.g. inferior frontal cortex)? Several studies suggest that it does not matter what “node” of the network (area of the brain) is stimulated; as long as some part of the network that is activated by the task is stimulated, activation of the entire network will be facilitated. If this result is found to be reliable, more expensive approaches to tDCS, such as high-definition tDCS or use of fMRI to locate the area to stimulate will be unnecessary.

The strongest evidence for pharmacological intervention is in the domains of post-stroke mood disorders and epilepsy, although they can also be useful in treating post-stroke pain and fatigue. Small RCTs indicate positive effects of SSRIs and other medications for improving recovery of cognition, language, and energy after stroke. Additional, multicenter RCTs are needed to replicate the findings to support the use of pharmacological interventions and noninvasive brain stimulation for non-motor symptoms. Furthermore, large, multicenter, randomized trials are required to identify the optimum timing, dose, and duration of medication use. To address these questions, sensitive measures of outcome will be essential. Trials that use modified Rankin Scale to evaluate outcomes are not sensitive to changes in motor function, but even less sensitive to changes in non-motor functions.

The influence of genetics on response to interventions is just beginning to be explored. For example, several studies have indicated that the BDNF phenotype influences response to treatment. (52, 98) Other allele, such as APOE alleles, and language related genes (e.g. FOXP2) might also influence response to treatment. Other factors, such as the health of the non-infarcted brain, indicated by the degree of white matter hyperintensities influences severity of deficits after stroke(99–101) and might also influence response to treatment.

The optimal timing of interventions after stroke remains hotly debated, more for motor rehabilitation than treatment of non-motor symptoms. Some authors argue that the first month or so after stroke, when neuroplasticity is greatest, provides a critical time period for rehabilitation to have a restorative, rather than a compensatory, effect(102). However, rehabilitation for chronic non-motor deficits has been shown to be effective in RCT as discussed above. It seems clear that clinicians should screen for non-motor symptoms as early as possible after stroke, to provide adequate counseling of families with or without rehabilitation (e.g. for aphasia, neglect, emotional communication deficits) as well as to prevent worsening (e.g. pain due to spactisity) or negative consequences for function (e.g. sleep and mood disorders).

Advances in imaging will also provide new insights on the effects of interventions on brain structure and function. Some studies have shown changes in neurotransmitters such as GABA, specific to the site of stimulation in tDCS trials.(103) Others have demonstrated changes in connectivity between the stimulated area and other cortical regions, using resting state functional MRI, obtained with tDCS(104). Still others report that changes in microstructural integrity or changes in functional connectivity are related to changes in functional gains(105). These changes in brain structure and function might be used as biomarkers of treatment response in future studies, as well as reveal some of the mechanisms underlying the effects of treatment.

Here I have reviewed recent trials for treatment of some, but not all, non-motor symptoms of stroke. I have not reviewed the explosion of studies on treatment of vascular cognitive impairment or vascular dementia, as potential treatments of those very common and disabling conditions requires a separate review.(81)

Despite its limitations this review highlights recent advances in treating many of the important non-motor symptoms of stroke. Most of the newer treatments, which are often adjuvants to rehabilitation, require larger, multicenter randomized placebo-controlled or sham-controlled trials to confirm their benefit. There is also a need to identify which treatments are best-suited to specific patients, based on genetics, imaging, or other biomarkers. Finally, future trials will need to show a benefit not only in terms of symptoms or impairments, but also in terms of function and quality of life.

Article highlights.

• Non-motor symptoms of stroke, including mood disorders, aphasia, hemispatial neglect, disorders of emotional communication, fatigue, seizures, sleep disorders, and pain, are common and disabling sequelae.

• Pharmacological interventions have been shown to be effective, through randomized clinical trials, for many post-stroke mood disorders, including depression and seizures, but are variably effective in treating post-stroke fatigue, sleep disorders, and pain.

• Rehabilitation, including physical and occupational therapy, is the primary intervention for hemispatial neglect and pain due to spasticity, but behavioral therapies might be augmented by non-invasive brain stimulation (or medications for spasticity).

• Speech-language treatment is the most important approach for treating aphasia and emotional communication deficits, but speech-language therapies might be augmented by non-invasive brain stimulation or medications.

• Sleep apnea is most effectively treated with Continuous Positive Airway Pressure (CPAP), although tolerance and compliance are limited (40–70%); alternative approaches should be tried when CPAP is not tolerated, due to the negative consequences of sleep apnea on outcome and stroke recurrence.