Fig. 14.1
Data on prevalence of post-stroke fatigue in selected large studies (>100 patients)
The onset of fatigue after stroke typically occurs early, during the acute hospital stay. In the South Korean study cited above, onset of post-stroke fatigue was within a week of stroke in 77 % of those affected, while only 10 % reported fatigue onset beyond 6 months [1]. In the Danish study that reported fatigue prevalence of 59 % at 10 days, only 9 % of patients developed fatigue beyond 3 months [2]. This is not to say that the acute stroke event is the trigger of fatigue in all cases. A number of people have already experienced fatigue before their stroke [1], and we return to the issue of pre-stroke fatigue in the discussion of contributing factors to post-stroke fatigue below. Nevertheless, it is likely that the causes of fatigue can be traced back to the stroke event itself, or to factors present in the first week post-event, or both.
It is notable that fatigue is persistent over time after stroke, despite marked improvements in neurological and physical impairments. Duncan and colleagues conducted a systematic review of longitudinal studies, in an attempt to map the natural history of fatigue after stroke [6]. They identified nine studies that assessed fatigue at multiple time points (up to 3 years post-stroke), finding that the frequency of fatigue declined over time in seven studies and increased over time in two studies. The overall picture, though, was that fatigue remained common even in the longer term. In the Danish study, fatigue was found in 59 % at 10 days, 44 % at 3 months, 38 % at 1 year, and 40 % at 2 years [2]. In the Dutch study, fatigue was identified in 52 % at baseline, 64 % at 6 months, and 70 % at 1 year [3].
The stroke event need not be severe for fatigue to manifest. In a study of 76 patients with minor stroke—independent in self-care and with no major cognitive impairments—56 % had fatigue at 6-month follow-up [7]. At this 6-month time point, the group was completely independent in activities of daily living, with a median score on the Barthel index [8] at the ceiling of 20. Another study included 99 functionally active, young (<70 years old) patients with a non-disabling first stroke (NIH Stroke Scale score <6). Of these patients, 35 % reported fatigue at 12-month follow-up [9]. In a similar cohort of younger, mild stroke survivors, 72 % experienced fatigue at 12 months post-stroke [10]. On the basis of these three studies, we can conclude that fatigue is experienced by the majority of mild stroke survivors.
To put these prevalence figures in context, it is important to consider the prevalence of fatigue in non-stroke control populations. Many studies have not included control samples, so the data here are more difficult to source. One comparison of 90 stroke survivors and 50 age-matched controls demonstrated that a significantly larger proportion of the stroke group than the control group (51 % versus 16 %) experienced severe fatigue [11]. The same fatigue prevalence of 51 % was identified in a younger group of ischaemic stroke patients (mean age 48), while fatigue prevalence in controls in this study was 32 % [12]. One well-designed study compared levels of fatigue in groups of patients with stroke, patients with chronic heart failure, and healthy controls [13]. Fatigue levels were similar in the stroke and heart failure groups. In adjusted multivariate analysis with the controls as the reference group, stroke patients were at six times greater risk (odds ratio = 6.18, 95 % CI 3.31–11.55) and heart failure patients were at eight times greater risk (odds ratio = 8.03, 95 % CI 4.63–13.94) for fatigue. The Danish study cited above included a reference group from the population; 32 % of these controls reported fatigue [2]. This group, however, was a mix of those with and without health complaints and was poorly matched to the stroke group. When the stroke group was compared with an age-, gender-, and living arrangement-matched subgroup of controls, fatigue was significantly higher in the patients at 10 days post-stroke (59 % versus 39 %), but the prevalence in the stroke group dropped back towards the control level over time. Together, these studies clearly indicate that stroke increases the likelihood of fatigue, but there is also a substantial number of people in the general population (perhaps around a quarter to a third of people) who experience fatigue.
Defining Post-Stroke Fatigue
In 2001, Staub and Bogousslavsky published a review titled “Fatigue after stroke: A major but neglected issue” [14]. Their paper brought much-needed attention to post-stroke fatigue. A remarkable feature of the review, and an indication of just how neglected the issue was, is the reference list of only 35 studies. Staub and Bogousslavsky defined subjective fatigue as “a feeling of early exhaustion developing during mental activity, with weariness, lack of energy and aversion to effort” [14]. Fatigue has been characterised as a state of weariness unrelated to previous exertion levels, and it is not usually overcome by rest [15]. We contend that the concept of fatigue should encompass physical as well as mental fatigue and not necessarily be brought on by periods of activity. To reflect this, a definition of fatigue borrowed from the multiple sclerosis (MS) literature is apt: “a subjective lack of physical or mental energy (or both) that is perceived by the individual to interfere with usual or desired activities” [16].
A systematic approach to developing a more formal case definition of post-stroke fatigue has been undertaken [17]. The definition reads: “Over the past month, there has been at least a 2-week period when the patient has experienced fatigue, a lack of energy, or an increased need to rest every day or nearly every day. This fatigue has led to difficulty taking part in everyday activities.” A similar, slightly modified definition was outlined for hospital inpatients. The structured interview that was formulated to classify fatigue emphasizes lack of energy and need to rest, and not lack of motivation, boredom, or sleepiness.
It is important to consider what fatigue is not. Fatigue should not be thought of as physical tiredness brought on by sustained exercise. A pertinent example of this in the stroke literature was provided by Tseng and colleagues [18]. They found that aerobic fitness was a strong predictor of the fatigue level reported immediately after exercise, whereas the strongest predictor of chronic fatigue was depression. “Exertion fatigue” may well be an important construct and is probably related to the broader notions of lack of physical and mental energy, but it is not what is meant when we refer to post-stroke fatigue. The same applies to mental fatigue: this should not be thought of simply as cognitive tiredness brought about by excessive mental effort.
Outside of stroke, some have drawn the distinction between pathological and non-pathological fatigue [19]. Fatigue is considered non-pathological if it is of short duration and has an identifiable cause (e.g., exercise, flu-like illness, endocrinopathy). Pathological fatigue has a greater intensity, longer duration and causes severe impairments to an individual’s functional ability and quality of life. This distinction can be usefully applied to the stroke setting: whether fatigue after stroke is considered pathological will depend on the timing and the impact on everyday life. Post-stroke fatigue can occur in the acute stage (which may be a normal, restorative response) or in the chronic stage (which is likely to be pathological). Pathological long-term fatigue is the greatest concern.
Another distinction that has been posited is the pathophysiological difference between central and peripheral fatigue [20, 21]. Central fatigue occurs when there is a failure to transmit motor impulses in the central nervous system, resulting in heightened perception of effort and reduced endurance in physical and mental activities. This failure has been localised to a network comprising the basal ganglia and its interconnected regions, including the thalamus and dorsolateral prefrontal cortex [22]. Peripheral fatigue is related to muscle fatigability and is characterised by failure to sustain the force of muscle contraction. The different roles of central and peripheral mechanisms underlying fatigue have not yet been explored in the context of stroke.
Measuring Post-Stroke Fatigue
Fatigue, as defined above, is a subjective experience, and therefore its measurement must include a phenomenological dimension. The most common approach to quantifying fatigue is via self-report assessment scales. Performance-based measures of fatigue can also be employed, with the realization that these measures reflect a narrower physiological conception of fatigue. For example, muscle fatigue can be evaluated by measuring the ability to perform muscle contractions over time [23]. Performance-based measures of mental fatigue have also been proposed, such as the use of a sustained attention task to assess cognitive fatigue [24]. In their review on fatigue in neurological conditions, Chaudhuri and Behan [20] outlined many techniques for identifying fatigability and its underlying causes. These included the Jolly test for identifying myasthenic disorders, single fibre electromyography, oxygen saturation of venous blood, muscle biopsy, neuroimaging to exclude demyelinating lesions, autonomic tests for orthostatic intolerance, and polysomnography. These physiological assessments offer attractive objectivity, but cannot be seen to encompass the concept of fatigue that was outlined earlier. The extent to which muscle fatigability and other physiological markers correlate with self-reported experience of fatigue after stroke is an interesting question that has not been addressed in the research literature.
Many of the self-report fatigue scales have been developed for and tested in MS. The similarities between these assessments are striking, both in item content and response scales, with only small differences in terms of focus on different aspects of fatigue. They are widely used in stroke populations, even though none have been developed specifically for this purpose (Table 14.1). One of the most commonly used measures in stroke is the fatigue severity scale [25]. It contains 9 items and responses to each item are made on a 7-point likert scale. The scale contains no specific reference to mental or cognitive fatigue; items are more focused on physical fatigue. A mean score >4 is considered to represent severe fatigue, though there is no strong rationale for this, and others have used ≥5 [26]. The fatigue assessment instrument is a predecessor of the fatigue severity scale, with a larger set of 29 items [27]. Another widely used tool is the fatigue assessment scale [28]. This 10-item measure uses a 5-point likert scale and covers both mental and physical aspects of fatigue. The fatigue impact scale is a 40-item tool, [29] with a modified version that contains 21 items on a 5-point likert scale. The shorter version consists of an 11-item cognitive subscale, a 7-item physical subscale, and a 3-item psychosocial subscale. The multi-dimensional fatigue inventory contains 20 items that are evenly distributed between the five dimensions of general fatigue, physical fatigue, mental fatigue, reduced motivation, and reduced activity [30]. Each item is scored from 1 to 5, with higher scores indicative of greater fatigue. Use of the total score is not recommended; those who want a single fatigue score are advised to use the general fatigue subscale score, and a cut-off of ≥12 on this subscale has been used to indicate fatigue [31]. A more recently developed tool is the fatigue scale for motor and cognitive functions [32]. The 20 items are evenly divided into mental and physical subscales, with responses on a 5-point likert scale. The checklist of individual strength also consists of 20 items, divided into the 4 subscales of subjective fatigue, concentration, motivation, and activity [33]. Suggested cut-offs are based on the subjective fatigue subscale, with <27 considered normal, 27–35 elevated, and >35 severe. While it is a broad health survey rather than a specific fatigue assessment, the Short Form-36 contains a vitality (or energy/fatigue) subscale of 4 items that is often used to measure fatigue [34]. Some have used the fatigue-inertia subscale from the Profile of Mood States assessment [35]. Brief visual analogue scales and single items have also been employed to identify fatigue, either in isolation (“Do you feel tired”) [36] or taken from other scales, such as depression screening tools [37].
Table 14.1
Self-report assessment scales for fatigue used in stroke populations
Assessment | No. of items | Dimensions | Original target population | Stroke studies |
|---|---|---|---|---|
Fatigue severity scale | 9 | – | MS, lupus | |
Fatigue assessment instrument | 29 | Global fatigue, situation-specific, consequences, responsive to rest | MS, chronic fatigue, Lyme disease, lupus, dysthymia | [9] |
Fatigue assessment scale | 10 | – | Workers | |
Fatigue impact scale | 40 (short-21) | Cognitive, physical, psychosocial | MS, chronic fatigue | |
Multidimensional fatigue inventory | 20 | General fatigue, physical fatigue, mental fatigue, reduced motivation, reduced activity | Cancer, chronic fatigue | |
Fatigue scale for motor and cognitive functions | 20 | Mental, physical | MS | [52] |
Checklist of individual strength | 24 | Subjective fatigue, concentration, motivation, physical activity | Chronic fatigue |
Assessing the criterion validity of these scales in stroke is not practical, as there is no accepted reference standard to evaluate them against. One attempt to evaluate the strengths and weaknesses of fatigue scales in stroke has been published: a comprehensive assessment of five common fatigue scales, chosen for their face validity in stroke [49]. Interestingly, the widely used fatigue severity scale was not one of them. Test-retest reliability (with a 3-day gap between assessments) was only moderate, with most scale items showing agreement between 0.40 and 0.60. Inter-rater reliability was generally very good, with agreement levels around 0.80–0.90. Attempting to move beyond reliability in evaluating fatigue scales is difficult. Demonstrating convergent validity of fatigue scales against each other only serves to illustrate what we already know—the items used are very similar. In terms of internal consistency, a lower Cronbach’s alpha is not necessarily a negative; this may simply reflect that a scale covers some of the broader aspects of the multidimensional concept that is fatigue.
The Importance of Post-stroke Fatigue
Stroke patients who are fatigued have lower quality of life than those without fatigue. Large studies in Norway [44], Hong Kong [45], and The Netherlands [47] have demonstrated that post-stroke fatigue is significantly related to quality of life, even after adjusting for age, disability, and depression. This is important as the quality of life is the closest thing we have to a central and universally important health-outcome measure. Quality of life incorporates not just physical, but also psychological and social domains of health and well-being [54, 55]. As healthcare has moved from the traditional medical model (with humans as biological organisms) to a more humane model (people as integrated, feeling beings), quality of life has become viewed as the endpoint most relevant to the individual [56]. It follows that any factor that has an independent effect on quality of life—such as post-stroke fatigue—constitutes an important target for intervention.
In a study of more than 4,000 Swedish stroke survivors, fatigue was independently associated with dependence in activities of daily living and with having to move into an institution following stroke [36]. Other studies have failed to identify an effect of fatigue on activities of daily living, even in the context of reduced quality of life [47]. This raises the prospect of simple self-care and household tasks that require low-level energy expenditure being unaffected by fatigue, but other activities that require more exertion and that are pivotal to quality of life being affected. Persistent fatigue documented at 2 years’ post-stroke was independently associated with a lower likelihood of returning to paid work (odds ratio = 0.29, 95 % CI 0.11–0.74).
Qualitative studies have shed light on the phenomenology of post-stroke fatigue. Fatigue matters to stroke patients, with 40 % reporting fatigue as their worst or one of their worst symptoms [50]. It limits participation in everyday life. One qualitative study reported that fatigue and the constant feeling of being tired (“It is just too much effort to do everything for myself”) was a primary reason for not engaging in activities [57]. In spite of reasonable objective recovery of physical function, fatigue in community-dwelling stroke survivors can be disabling [58]. Even in patients with low levels of physical disability, fatigue has debilitating effects on social participation, return to work, driving, reading, and sleeping [59]. Many patients reported feeling unprepared for the fatigue and struggling to adapt to it. Looking at the impacts reported here, it is easy to see how fatigue may influence quality of life (e.g., less social interaction, limited employment opportunities) without having a major effect on independence in activities of daily life.
It is likely that fatigue can have profound effects on stroke rehabilitation and inability to fully participate in rehabilitation may be a mediating factor behind many of the deleterious effects of fatigue. Unfortunately, there are very few data on this issue. One review was titled “Fatigue associated with stroke and other neurologic conditions: Implications for stroke rehabilitation,” [15] yet this paper did not identify any studies on the effect of post-stroke fatigue on rehabilitation. Michael [60] suggested that fatigue can impede participation in rehabilitation, but this paper was based on anecdotal evidence from a single clinical case and a discussion of fatigue models, rather than results from an original data set. Morley and colleagues addressed the question in a pilot study of 20 stroke rehabilitation inpatients, with data reported briefly in a letter to the editor [61]. Eight of the 20 patients were assessed as having fatigue on the fatigue severity scale. Information from physiotherapists on whether fatigue interfered with rehabilitation was available for 16 patients; it had interfered in 6 of the 16, but only 3 of these 6 were fatigued according to the scale. So while it makes intuitive sense that fatigue can hamper rehabilitation after stroke, and therefore impact on recovery, there is very little supporting evidence for this relationship.
Several studies have raised the prospect that fatigue after stroke is related to an increase in mortality. Three years after stroke, patients with fatigue had a higher case fatality rate than those without fatigue [36]. In a UK study of more than 1,000 stroke patients, presence of fatigue was associated with shorter subsequent survival [62]. In a group of young stroke patients (mean age 48), Naess and colleagues reported that, adjusting for age and sex, fatigue was associated with mortality across the 12-year follow-up [43]. There is also evidence that post-stroke fatigue confers a higher risk of suicidality [46]. Therefore, reduction in levels of fatigue may result in not only better quality of life, but also longer life for stroke survivors. These findings do not necessarily imply that fatigue is a direct cause of death; it is likely that the fatigue-mortality relationship is mediated by other factors. In the Naess study, the significant relationship between fatigue and mortality disappeared when the multivariate regression included not only age and sex, but all variables associated with mortality in univariate analyses (alcoholism, myocardial infarction, and unemployment).
Potential Causes of Post-stroke Fatigue
As fatigue is a major contributor to the burden of disease following stroke, we need to understand what factors predispose patients to ongoing fatigue. There is not a long history of investigation into post-stroke fatigue, and we still have very limited understanding of the mechanisms behind it. Of the studies that do exist, most have been observational and cross-sectional. We might yearn to identify causal factors for fatigue, but the bulk of our current information is at the level of association, not causation. There are published reviews of post-stroke fatigue that have made valuable contributions to summarising the existing literature [63, 64]. In the following discussion, we will split the variables that are potentially associated with the development of fatigue into three areas: pre-existing and stroke-specific factors, co-existing and bi-directional factors, and modifiable behavioural and environmental factors.
Pre-existing and Stroke-Specific Factors
Demographic Factors
There are conflicting results regarding the influence of age and sex on post-stroke fatigue. Older age was related to a greater likelihood of fatigue at 1 year [3] and at 15 months [16] post-stroke. Conversely, other studies have found that younger age is linked to a greater likelihood of fatigue at 2 months [65] and at 1 year [51] post-stroke. Completing this equivocal picture are studies that failed to identify a significant association between age and fatigue [1, 12, 50]. It is difficult to account for the difference between these studies and draw firm conclusions about the relationship between age and post-stroke fatigue. Some may assume that the likelihood of experiencing fatigue will increase as people get older, but this is not borne out in the data.
Studies in the general population have indicated that fatigue prevalence is higher in women than men [66, 67]. Interestingly, one stroke study found this sex difference in their control participants, but did not identify any relationship between sex and fatigue in stroke survivors [50]. There are other studies that report no sex differences in fatigue after stroke [1, 11], but these are outweighed by the number of studies that have identified higher levels of fatigue in women than men after stroke [2, 3, 5, 36, 39, 62]. Women are also more likely to be depressed than men, but the gender imbalance in fatigue was present in several of these studies even after accounting for depression. The imbalance is not necessarily one of biology and physiology: another possibility is that female stroke survivors receive less support in completing household tasks than male stroke survivors and are thus more prone to fatigue.
Pre-Stroke Fatigue and Vascular Burden
Information on pre-stroke fatigue is typically acquired retrospectively, so reliability of the data might be called into question. With that caveat, there is evidence that pre-stroke fatigue is the most important factor in explaining post-stroke fatigue, beyond even depression and functional independence [1]. Another study reported that pre-stroke fatigue was independently related to fatigue in the acute stage of stroke [39]. Given the vascular compromise (e.g., small-vessel disease, subclinical stroke) and co-morbidities (e.g., diabetes, heart failure) that exist prior to stroke, it is possible that pre-stroke fatigue may be a marker for these conditions. The findings that leukoaraiosis is independently associated with post-stroke fatigue [26], and that patients with pre-stroke fatigue have more co-morbidities than those without pre-stroke fatigue [1], are consistent with this argument.
Stroke-Specific “Organic” Factors
Organic factors stemming from the brain lesion itself may play an important role: fatigue is more prevalent after minor stroke (56 %) than after transient ischemic attack (29 %) [7]. These rates of fatigue were barely altered when stroke patients with baseline NIHSS of 0 were compared to TIA patients (57 % versus 29 %). The argument follows that fatigue is not simply a consequence of a stressful acute cerebral event, comorbidity, medication, or level of disability, as these factors were all comparable between the groups. If this is true, we should be able to identify some properties of stroke-related damage that are associated with the development of fatigue.
Lesion Location
It has been suggested that damage to the brainstem and reticular formation is likely to predispose to fatigue [14]. This was supported by the finding that infratentorial infarctions were related to an increased risk of fatigue (odds ratio 4.69, 95 % CI 1.03–21.47) [65]. We have known for 100 years that lesions in the brainstem can cause specific alterations to wake-sleep regulation [68]. Given the anatomy of the ascending reticular activating system, it is unsurprising that these brainstem lesions can block ascending pathways and produce a profound impairment of arousal that could be experienced as fatigue. It is clear, though, that the genesis of post-stroke fatigue cannot be reduced to brainstem damage alone. One study that included only patients with supratentorial stroke identified a high fatigue prevalence of 70 % at 1 year post-stroke [3]. Another study found that basal ganglia infarcts are more likely to produce fatigue than brainstem infarcts [5]. Yet many studies have failed to identify an association between lesion location and fatigue [63]. The best way to approach this question is to synthesise all the current evidence. In 2012, a systematic review was published on the link between lesion location and fatigue [69]. There was no conclusive evidence of a relationship; meta-analysis was not conducted due to the methodological differences between studies. Four (n = 675) of 13 studies (n = 1613) reported associations between fatigue and infratentorial lesion location (particularly brainstem) or basal ganglia stroke. It should be remembered that the absence of evidence does not mean evidence of absence. Simple structural metrics such as location and size of the lesion may not reflect the most important aspects of damage; we also need to consider interruptions to large-scale neural networks, perfusion abnormalities, changes to functional connectivity, and other alterations that may be related to fatigue.
Stroke Type
Surprisingly little attention has been paid to whether stroke subtype has a differential effect on fatigue. Haemorrhagic stroke tends to be more severe than ischaemic stroke and is associated with increased mortality risk in the acute stage [70]. It might be argued that the haemorrhagic nature of the lesion could predispose to fatigue. When ischaemic and haemorrhagic stroke patients have been compared, however, no difference in occurrence of fatigue has been identified [3]. A study that had a more detailed breakdown of subtype yielded the same finding: there were no differences in fatigue between patients with intracerebral haemorrhage, large vessel infarct, lacunar infarct, or embolic infarct [71].
Neural Activity and Metabolism
In addition to structural alterations, brain imaging techniques have been used to investigate functional, metabolic, and perfusion changes in fatigue, although generally not in the context of stroke. Abnormal cortical activation on EEG has been associated with fatigue in both MS [72] and postpoliomyelitis [73] patients. Functional MRI has been used to show reduced functional interaction between cortical and subcortical areas in fatigued MS patients [74]. FDG-PET scanning has been employed in fatigued MS patients, revealing areas of hypometabolism in prefrontal cortex and adjacent white matter, premotor, and supplementary motor areas and putamen [75]. These authors hypothesised that fatigue in MS is associated with frontal cortex and basal ganglia dysfunction that might result from white matter demyelination. Whether similar changes in brain activity occur in post-stroke fatigue is yet to be documented.
Inflammatory Markers
There is a strong rationale to predict that pro-inflammatory cytokines are related to higher levels of fatigue after stroke. Patients with chronic fatigue syndrome often have elevated C-reactive protein and other inflammatory markers [76]. Inflammation has a prominent role in the ischaemic cascade that is triggered by stroke and might help to explain the onset of fatigue. This is another area, however, where data are scarce. One study quantified cytokines and other blood components in acute stroke and investigated their relationship to fatigue in the longer term [77]. Higher levels of glucose and IL-1β were associated with greater fatigue at 6 months, and higher levels of glucose and lower levels of IL-1ra and IL-9 were associated with greater fatigue at 12 months. The authors argued that cytokine-induced sickness behaviour, especially in relation to IL-1β, could explain symptoms of fatigue, with disturbed glutamate signalling the most plausible mechanism. The data require replication, however, especially in light of the numerous comparisons made and the fact that none of the measured biomarkers were associated with fatigue at 18 months. One other study has investigated C-reactive protein, but this was only a small pilot in 28 patients [78]. There was some indication of higher levels of C-reactive protein in the fatigued group, but this could only be detected after excluding those with pre-stroke fatigue or symptoms of mood disorder.
One argument that has been proposed is that fatigue is associated with an underactive hypothalamic-pituitary-adrenal (HPA) axis. The HPA axis forms a major part of the neuroendocrine system, controlling hormone release through complex feedback mechanisms in reaction to stress, and to regulate many other bodily functions. There is evidence that the HPA axis is underactive in patients with chronic fatigue syndrome, fibromyalgia, and post-poliomyelitis fatigue [20]. Pre-existing low cortisol levels might sensitise the HPA axis to the development of central fatigue after an acute stressor (such as stroke). Cortisol secreted by the adrenal gland acts via a negative feedback mechanism to reduce hormone release from the hypothalamus and anterior pituitary. If cortisol levels fail to normalise in the chronic phase of stroke, continued down-regulation of the HPA axis may result in persistent fatigue.
Comorbidities
Stroke patients often have numerous comorbidities, and these comorbidities may be important factors in the development of fatigue. The common stroke-related co-morbidities of diabetes mellitus and myocardial infarction were both independently related to fatigue in a sample of 377 stroke patients [26]. These same two co-morbidities were found to be significantly associated with fatigue in a young cohort of stroke survivors, though the correlations were not strong (r = 0.13 and 0.16) [43]. We know that the presence of diabetes and myocardial infarction are linked to mortality, and as such are possible mediators of the fatigue-mortality association. In a multivariate analysis of stroke patients and controls together, greater fatigue severity was independently associated with diabetes mellitus, alongside cerebral infarction, depression, and migraine [12]. This is interesting as it suggests that the effects of diabetes, depression, and migraine on fatigue are at least partially separable from the effects of stroke. The contribution of diabetes and ischaemic heart disease is not completely consistent: large studies exist that did not find higher levels of fatigue in the context of these comorbidities [1, 71]. These studies also reported similar rates of hypertension in those with and without post-stroke fatigue. The impact of blood pressure is complex, with the relationship to fatigue likely to be U-shaped rather than linear; that is, both hypertension and hypotension may lead to greater fatigue. We know that hypotension is related to fatigue in the general population [79]. In a detailed study of ambulatory blood pressure monitoring in stroke patients, fatigue was more common in those with either hypertension or hypotension [80]. An added complication in stroke populations is the question of attribution: fatigue may result from the direct effects of low blood pressure, but it may also be explained by the use of antihypertensive medications, many of which are known to have symptoms of fatigue as side effects [80].
Medications
It is not only antihypertensives that have fatiguing effects. Antidepressants, anxiolytics, and benzodiazepines can all have sedative effects, producing drowsiness and lethargy that may manifest as fatigue. This is particularly relevant given the high prevalence of mood disorders and sleeping problems after stroke, increasing the likelihood that these medications will be prescribed. There is also evidence that another class of medication often used after stroke—statins—can sometimes produce rhabdomyolysis, and this muscle breakdown can be experienced as malaise and fatigue [81].
Pain
Stroke survivors who experience pain are particularly vulnerable to physical inactivity, poor sleep, and mood disorder, all factors that can contribute to fatigue. Naess and colleagues found that pain was significantly associated with post-stroke fatigue, independent of depression, and sleep disturbance [26]. Another study investigated the associates of both pain and fatigue after stroke, but did not assess the relationship between these two factors [71].
Co-existing and Bi-directional Factors
Physical Disability
Studies have identified strong and independent relationships between post-stroke fatigue and higher levels of disability [1, 12], but it is difficult to draw firm conclusions about the direction of causality. Profound one-sided weakness is a common acute stroke symptom, and poorer physical function in the first 2 weeks after stroke has been related to fatigue [39]. In a study that included both stroke patients and controls, the factor that explained most of the variance in fatigue in the stroke group was impairment in locomotion, whereas depression was the predominant factor in accounting for fatigue in controls [11]. Not only does hemiparesis contribute to physical inactivity, which can exacerbate fatigue, the compensatory strategies and motor re-learning that is required may also increase fatigue. The inefficiency of hemiparetic gait necessitates higher-than-usual energy expenditure; there is evidence that the energy cost of walking following stroke is double that of controls [82]. The relationship is potentially bi-directional because the experience of fatigue can limit participation in rehabilitation and reduce task practice with the affected limbs, thus feeding into continued physical disability. Of course, post-stroke physical disability is a broader concept than hemiparesis. One stroke study identified dysarthria, a motor disorder of speech production, in 30 % of those with fatigue but only 14 % of those without fatigue [1]. Another study considered the items from the baseline NIHSS separately and found that facial palsy and arm paresis were significantly associated with 12-month fatigue; in this analysis dysarthria was not related to fatigue [71].
Depression
Of all the factors that have been associated with post-stroke fatigue, depression is the one that stands out as the most prominent. The finding of a significant independent association between depression and fatigue after stroke is well documented [3, 12, 44, 65]. In a meta-analysis of 19 stroke studies (n = 6,712), depression and fatigue were significantly associated (OR = 4.14, 95 % CI 2.73–6.27) (Fig. 14.2) [83]. The relationship has been identified in the first 2 weeks after stroke [39] and also at several years following stroke [4]. It is not only severe strokes that are linked to mood disorder and fatigue. In a group of people who were under age 70 and had experienced small, non-disabling first strokes, depression was associated with fatigue across the 12 months of follow-up [9]. Interestingly, there is some evidence to indicate that, controlling for other factors, a stroke patient is more likely to experience fatigue if they were depressed prior to their stroke [26].
