Key points
- •
Motor function at baseline and age are the strongest clinical predictors of stroke recovery. Bedside examination, such as Medical Research Counsel testing, can provide the clinician with enough information to predict long-term recovery.
- •
Ambulation ability is a strong predictor of long-term independence. Ability to regain walking ability can be predicted by balance ability at onset.
- •
Arm and hand recovery can be predicted by motor function at onset, but actual use of hand in functional activities requires significant recovery of motor speed and manipulative skill.
- •
Language recovery can be predicted by language ability at stroke onset. A good ability to comprehend language and intact repetition predict better language outcome.
- •
Combining motor ability, neurophysiological measures such as preservation of motor-evoked potentials by transcranial magnetic stimulation, and imaging may provide useful predictive value if current available models can be validated in large patient cohorts.
Introduction
Accurate prediction of functional outcome in patients with stroke has the potential to enhance clinical care as well as improve the quality of stroke research. Prognostic models can facilitate education and counseling of patients and families, and streamline planning for rehabilitation and discharge. Specific predictors can help target treatment options to the patients who will most benefit, and avoid treatments in those who are unlikely to respond. Such models can improve research analysis when adjusting for baseline characteristics in study cohorts and comparing different randomized trials for meta-analysis.
The use of formal prognostic models to predict functional outcome have not been used in clinical stroke rehabilitation because large representative cohorts have not been studied and existing models are not well validated. Thus, prognosis has been limited to a handful of “clinical pearls” and the clinician’s personal experience, with the help of existing literature. It is important to note that the literature on prognosis has provided some useful information on recovery of impairment and activity, but fewer data on participation. Additionally, there is a paucity of research on the impact of cognitive and perceptual dysfunction and recovery. Most of the available literature is composed of small cohort studies and systematic reviews. Although formalized and validated predictive models would be an improvement over current practice, there remains much value in the use of “rules of thumb” for prognosis of functional outcome. The purpose of this review was to attempt to list those rules of thumb based on existing prognostic models, current epidemiologic evidence, and my own experience as a stroke rehabilitation specialist.
Introduction
Accurate prediction of functional outcome in patients with stroke has the potential to enhance clinical care as well as improve the quality of stroke research. Prognostic models can facilitate education and counseling of patients and families, and streamline planning for rehabilitation and discharge. Specific predictors can help target treatment options to the patients who will most benefit, and avoid treatments in those who are unlikely to respond. Such models can improve research analysis when adjusting for baseline characteristics in study cohorts and comparing different randomized trials for meta-analysis.
The use of formal prognostic models to predict functional outcome have not been used in clinical stroke rehabilitation because large representative cohorts have not been studied and existing models are not well validated. Thus, prognosis has been limited to a handful of “clinical pearls” and the clinician’s personal experience, with the help of existing literature. It is important to note that the literature on prognosis has provided some useful information on recovery of impairment and activity, but fewer data on participation. Additionally, there is a paucity of research on the impact of cognitive and perceptual dysfunction and recovery. Most of the available literature is composed of small cohort studies and systematic reviews. Although formalized and validated predictive models would be an improvement over current practice, there remains much value in the use of “rules of thumb” for prognosis of functional outcome. The purpose of this review was to attempt to list those rules of thumb based on existing prognostic models, current epidemiologic evidence, and my own experience as a stroke rehabilitation specialist.
Patterns of recovery from stroke and key measures of outcome
Motor recovery has been extensively studied in stroke due to the availability of reliable and valid measures. The first rules of thumb were provided by Twitchell in 1951 when he described the patterns of natural recovery from stroke. In observing a cohort of stroke survivors he concluded the following sequential processes of recovery:
- •
Initial loss of voluntary movement and reflexes
- •
Rapid restoration of reflexes proceeding to hyperreflexia
- •
Development of increased muscle tone
- •
First voluntary movements in shoulder and hip
- •
Appearance of further voluntary movement with flexor pattern in upper limb and extensor pattern in lower limb
- •
Both flexor and extensor movements appear in upper and lower limbs
- •
Spasticity is reduced as isolated joint and finger movements emerge
Twitchell noted that patients can progress quite quickly through this recovery pattern or stop recovering at any given level depending on stroke severity. Signe Brunnstrom used Twitchell’s findings to develop a scale of motor impairment after stroke. Later, Axel Fugl-Meyer and colleagues used Twitchell’s principles to design a more detailed scale of motor impairment using an elegant scoring system. The upper limb portion of the Fugl-Meyer assessment (FMA) is now considered a standard outcome measure in clinical stroke recovery research.
Although the lower limb portion of the FMA also has been used in stroke research, recent clinical trials have used walking speed as an ideal measure of lower limb recovery. Walking speed is associated with lower limb impairment, aerobic capacity, and functional ambulation. Improvements in walking speed also correlate with improved overall ability, as measured by the Modified Rankin Scale ( Table 1 ).
| 0 | No symptoms. |
| 1 | No significant disability. Able to carry out all usual activities, despite some symptoms. |
| 2 | Slight disability. Able to look after own affairs without assistance, but unable to carry out all previous activities. |
| 3 | Moderate disability. Requires some help, but able to walk unassisted. |
| 4 | Moderately severe disability. Unable to attend to own bodily needs without assistance, and unable to walk unassisted. |
| 5 | Severe disability. Requires constant nursing care and attention, bedridden, incontinent. |
| 6 | Dead. |
Although it is a fairly insensitive measure of functional activity, the Modified Rankin Scale (MRS) is a standard measure of outcome in acute stroke research and has been used in studies on early prediction of outcome. Along with ambulation, overall severity of motor function is related to MRS, with minimal motor impairment being associated with favorable outcome, typically defined as an MRS ≤2 (slight disability or better). But it is important to recognize that a favorable outcome on the MRS is not strongly related to recovery of the affected arm and hand. Aphasia, on the other hand, is associated with greater dependence as measured by MRS. This is not surprising, given that significant residual deficits in comprehension predict a lower probability of return home following acute rehabilitation and are associated with lower motor and cognitive scores on the Functional Independence Measure (FIM).
Thus, severity of impairment after stroke is related to overall functional ability. Better motor recovery in the arm and leg, faster walking speed, and good language comprehension result in greater long-term independence. With these principles as a foundation, the current literature on prediction of functional outcome following stroke is reviewed. The focus is primarily on clinical factors that are predictive of stroke outcome, but the more recent use of imaging and neurophysiological measures also is briefly discussed.
Predictors of functional outcome after stroke
Following Twitchell’s description of natural motor recovery in stroke, he described factors that were positive and negative predictors of outcome:
- •
Predictors of better recovery
- ○
Only mild spasticity at its worst and none at shoulder
- ○
Rapid progression through synergy to isolated movement
- ○
- •
Predictors of worse recovery
- ○
Late return of reflexes
- ○
Late onset of voluntary movement
- ○
Increasing severity of spasticity
- ○
In practice, these principles seem to hold true. From here, research on outcome for general recovery of activities of daily living, walking ability, upper limb skill, and language function is presented.
Activities of Daily Living
Most studies examining overall recovery of independence in activities of daily living (ADLs) use the Barthel Index (BI) or the MRS. The clinical measurements that are strongly associated with long-term (>3 months) independence in ADLs are age, the National Institutes of Health stroke scale (NIHSS) and the early BI score. Younger age consistently predicts favorable ADL outcome, defined as an MRS ≤2 or a BI ≥95. A lower score on NIHSS at admission (<10) predicts a favorable outcome; however, the closer the NIHSS is to zero, the better the prediction of independence. Higher scores on BI measured early after stroke are associated with favorable outcomes. But unlike the NIHSS, which has a good predictive strength at 48 hours after stroke, the BI is a better predictor of long-term ADL when measured at 5 days or later. The unreliability of BI within the first 48 hours is likely due to the challenge of accurately determining functional ability early after stroke admission.
Male gender has been associated with better functional outcome in several studies, but in a large systematic review, gender was found to have no relation to ADL outcome. Other less consistently reported clinical predictors of long-term ADL independence include urinary continence, good sitting balance, absence of aphasia, and absence of diabetes mellitus. Diabetic patients, especially those who are older, face a higher risk of severe stroke leading to worse outcome. No other medical comorbidities have been associated with functional outcome. Jongbloed noted in a systematic review that longer time before hospital admission was associated with worse functional ability at discharge, suggesting that earlier hospital arrival provided for more effective acute treatment. In contrast, Tei and associates found that longer time before hospital admission was associated with favorable outcome at 3 months after stroke, suggesting that persons with very mild stroke symptoms are less likely to initiate timely emergency care. Recently, Ntaios and colleagues developed the ASTRAL score, which uses patient age, NIHSS, time from stroke to admission, visual field deficit, admission glucose, and level of consciousness on admission to predict mortality and poor outcome (>2 on MRS) at 3 months or later ( Table 2 ). This score has been validated in several populations.
| Clinical Factors | Points |
|---|---|
| Age: for every 5 y (from 0) | 1 |
| Severity: for every NIHSS point | 1 |
| Time delay: time from onset >3 h | 2 |
| Any visual field deficit | 2 |
| Acute glucose >7.3 or <3.7 mmols/L | 1 |
| Level of consciousness decreased on NIHSS | 3 |
There is an expanding use of cranial imaging for the prediction functional outcome after stroke. But it is not yet clear whether prognosis based on imaging can exceed that of clinical assessment. For example, larger stroke volume on anatomic imaging is associated with worse functional outcome, but early motor strength and BI have stronger predictive value. Stroke volume measured on acute diffusion-weighted imaging (DWI) is not superior to NIHSS or age in predicting late MRS. However, Zaidi and colleagues found that small final infarct volume on DWI following thrombolysis in a cohort of patients with middle cerebral artery stroke predicted better MRS score (<2) at 90 days after stroke. Very large strokes result in poor outcome, but the impact of small-volume and moderate-volume strokes is dependent on the structures involved. For example, strokes that injure the posterior limb of internal capsule (PLIC) predict motor outcome better than overall stroke volume.
Analysis of cortical spinal tract (CST) structure using diffusion tensor imaging (DTI) and tractography improves prediction of outcome. DTI is a magnetic resonance technique that measures the directional movement of water molecules along white matter tracts and is an anatomic measure of tract connectivity. Measures of the water flow along a tract (anisotropy) versus in no particular direction (isotropy) are useful for differentiating intact versus injured portions of white matter, respectively. Using DTI after stroke, Radlinska and colleagues showed that a lower tract volume and lower fractional anisotropy (FA) were associated with worse functional performance at 6-month follow-up.
In conclusion, neurologic status, motor ability, and function at stroke onset best predict long-term outcome on ADLs. Younger age predicts better outcome, whereas injury to PLIC on cranial imaging is associated with worse outcome.
Ambulation
Prognosis for ambulation after stroke is fairly good, such that 70% to 80% of chronic stroke survivors have the ability to walk. On the negative side, only approximately 30% to 50% return to community ambulation, with gait speed being a key determinant of success. There have been no systematic reviews to identify predictors for walking after stroke. Two small studies have shown that sitting balance 2 weeks after stroke onset can predict 6-month ambulation ability. In particular, a score of ≤50 on the Trunk Control Test 14 days after stroke predicts that walking is unlikely at 6-month follow-up ( Table 3 ). Better trunk control on admission to rehabilitation also predicts better discharge motor FIM and better ADL performance 6 month after stroke. Age, NIHSS, baseline FIM score, and motor strength are also correlated with walking outcome, but none of these are as strong a predictor as balance. Although early sitting balance can predict later walking, only gains in dynamic (walking) balance are associated with improvements in long-distance ambulation.
| Activity |
|---|
| Rolling to weak side |
| Rolling to strong side |
| Sitting up from lying down Balance in Sitting position |
| Scoring for Each Activity |
|---|
| 0 = Unable to do on own |
| 12 = Able to do but only by pulling on rail bedsheets and or stabilizing on arm |
| 25 = Normal |
| Total = 100 points |
Reding and Potes plotted life tables showing that probability of walking after stroke is related to stroke severity. In particular they showed that more than 90% of patients with stroke will walk again, with or without an assistive device or some assistance after stroke. This goal was achieved for patients with pure motor hemiplegia at approximately 14 weeks after stroke, at 22 weeks for those with motor and sensory deficits, and at 28 weeks for those with motor, sensory, and visual deficits. Thus, it is not whether patients can walk with or without assistance after stroke, but when. When considering fully independent ambulation, the rates of recovery were as follows: 90% of those with only motor deficits are independent walkers at 14 weeks after stroke, 35% with motor and sensory deficits walk at 22 weeks, and only 3% of those with motor, sensory, and visual deficits are walking on their own even at 30 weeks after stroke.
As mentioned previously, there is a clear relationship between gait speed and overall functional independence. Both higher gait speed and timed distance walking are equally predictive of community ambulation after stroke. Higher gait speeds are related to better aerobic endurance, leg strength, balance, and a low fear of falling. Although studies assessing early predictors of later community ambulation after stroke are lacking, one small study showed that early (<3 months) ability to walk at a speed of just over 0.4 m/s, and a low fear of falling, can predict community ambulation at 6 months. These associations suggest, as has recent clinical research, that high-intensity gait training at aerobic levels of effort, focusing on walking speed, might improve timed gait distance and potentially facilitate return to community ambulation.
There has been limited research on predicting ambulation outcome with imaging. Lesion size and location have been associated with gait severity, but the level of injury to CST after stroke does not predict later walking speed or response to gait training. There has recently been a growing interest in the use transcranial magnetic stimulation (TMS) for predicting motor outcomes because of its ability to measure the integrity (or conductivity) of motor systems in the central nervous system. The absence of TMS-induced motor-evoked potentials (MEPs) in the tibialis anterior muscle of a hemiplegic limb is associated with poor recovery of transfer ability and walking, but the predictive value TMS for functional recovery in lower limb has not been verified.
Upper Limb Dexterity
The recovery of motor control and function in the upper limb follows a similar rate and pattern as the lower limb using standardized tests of both impairment and activity. But actual functional use of the upper limb after stroke demands the return of a high degree of fine motor skill. For example, Flemming and colleagues found in a small cohort of stroke survivors that a dexterity score of ≥54 of the maximum 57 on the Action Research Arm Test (ARAT) was necessary for patients to report that they use their impaired upper limb half as much or better as before stroke. The amount of recovered hand skill has also been shown to have a strong impact on patients’ reported health-related quality of life. In 2001, Coupar and others reported findings from a meta-analysis that included 58 studies examining recovery of upper limb function after stroke. Not surprisingly, this analysis was hampered by the large variety of outcome measures used to assess performance. Still, they were able to conclude that the only clinical predictors that were strongly associated with upper limb recovery were baseline arm and hand motor ability and function. Lower motor performance at baseline results in a lower arm and hand ability 3 to 12 months after stroke. These findings were consistent with a previous systematic review by Chen and Winstein. The measures used for baseline motor function in these studies included the Medical Research Council (MRC) strength testing, motricity index, NIHSS arm motor score, Scandinavian stroke scale arm impairment score, FMA, ARAT, Rivermeed arm score, BI, the motor assessment scale and others. Limited upper limb recovery was also associated with baseline lower limb motor function, but to a lesser degree. Thus, even simple bedside tests, such as manual muscle grade (MRC) or motricity index, which is composite score of the MRC for both upper and lower limb, are the simplest tools clinicians can use to predict later arm recovery ( Table 4 ). Au-Yeung and Hui-Chan determined that a motricity index score of 64 of 100 or better in the impaired upper limb at 4 weeks after stroke predicts an ARAT of ≥35 at 6 months. If 2-point discrimination on the affected limb is normal at 2 weeks, then an upper limb motricity index of 45 or better predicts the same outcome. Of note, younger patients with low motor scores at baseline have the potential to recover hand dexterity even beyond 6 months after stroke. Sensorimotor testing likely has its limitations. Strength does predict better outcome on standardized tests of upper limb dexterity and function, but even a detailed sensorimotor assessment of impaired upper limb explains only 25% of the variance when measuring the quality of a reaching task, such as speed, accuracy, and efficiency. Thus, there is far more involved in recovering skilled movement than motor strength alone.
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
