Keywords:
Stroke - Heart Rate - Rehabilitation - Physical Therapy
Palavras-chave:
Acidente Vascular Cerebral - Variabilidade Cardíaca - Reabilitação - Fisioterapia
INTRODUCTION
Stroke is one of the main causes of morbidity and mortality in industrialized countries
and the leading cause of chronic disability in adults[1]
-
[3]. After stroke, more than 70% of individuals present alterations in motor, sensory,
or cognitive systems, which can be mild and transient or severe and disabling, and
these alterations can be related to autonomic nervous system impairments, which can
lead to changes in heart rate variability (HRV)[4]
-
[6].
HRV is the result of adaptive changes in heart rate caused by sympathetic and parasympathetic
activity in response to external or internal stimuli[7]. Based on this concept, HRV is defined as the changes in heart rate (HR) that occur
after a stimulus, and it is a predictor of processes related to the autonomic nervous
system. Studies have shown that a low HRV response is related to a high risk of stroke[8],[9], severe stroke severity[10], mortality after stroke[4],[5],[11], low vagal modulation[12], and a poor prognosis after stroke[13].
There is evidence that physical inactivity reduces cardiac autonomic modulation after
stroke[14]. Therefore, the autonomic nervous system can be increased through physical exercise
and rehabilitation programs after stroke[15]. Lower HRV is a predictor of morbidity and mortality and cardiac changes increase
the risk of death after stroke[16] and may be related to unfavorable outcomes[17]. Additional studies need to be conducted to elucidate the cardiac autonomic modulating
mechanisms and clinical repercussions of HRV after stroke rehabilitation.
Thus, it is possible that specific and effective rehabilitation programs, allowing
greater cardiovascular stability, functional gains, and quality of life in individuals
after stroke, can be developed. Due to the lack of evidence that rehabilitation can
be effective in modulating the autonomic nervous system after stroke, there is no
consensus on this effect; there are no systematic reviews in the literature on this
topic. Therefore, the aim of this review was to evaluate the effectiveness and safety
of rehabilitation programs in modulating HRV after stroke.
METHODS
We adhered to the methods described in the Cochrane Handbook for Intervention Reviews[18] and to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA)
reporting guidelines[19]. This review was registered in the International Prospective Register of Systematic
Reviews (PROSPERO - CRD42020156527).
Eligibility criteria
The eligibility criteria were as follows:
-
Study designs: RCTs, quasi-RCTs, and non-RCTs
-
Participants: adults over 18 years of age of either sex with any duration of illness,
severity of initial impairment, type of stroke diagnosis (ischemic or intracranial
hemorrhage) that was made by a clinical examination or radiographically by computed
tomography (CT) or magnetic resonance imaging (MRI).
-
Interventions: any rehabilitation protocol for stroke recovery (early mobilization,
physical exercises)
-
Comparators: any conventional stroke rehabilitation program
-
Outcomes: Heart rate variability
Data sources and search strategy
The search strategy was based in the PICOT (patients: stroke; intervention: rehabilitation;
comparison: any control group; outcome: heart rate variability; time: acute, subacute,
and chronic phases of stroke). We searched MEDLINE (OvidSP), the Cochrane Central
Register of Controlled Trials (CENTRAL), CINAHL, the Latin-American and Caribbean
Center on Health Sciences Information (LILACS), and SCIELO databases without language
restrictions. The date of the most recent search was July 10, 2020. All searches were
conducted with the assistance of a trained medical librarian. We also searched the
reference lists of relevant articles and conference proceedings, and contacted the
authors of the included trials.
The search terms included “Heart rate variability or (MeSH terms)” and stroke or (MeSH
terms) and rehabilitation or (MeSH terms).
Other resources searched
In an effort to identify additional published, unpublished, and ongoing trials, we
performed the following steps:
-
screened the reference lists of the identified studies;
-
contacted the study authors and experts; and
-
used the Science Citation Index Cited Reference Search to track important articles.
Selection of the studies
Two pairs of reviewers independently screened all titles and abstracts identified
in the literature search, obtained full-text articles of all the potentially eligible
studies, and evaluated the articles for eligibility. The reviewers resolved disagreements
by discussion or, if necessary, with third party adjudication. We also considered
studies reported only as conference abstracts.
We used the START program (State of the Art through Systematic Review), developed
by the Software Engineering Research Laboratory of the Federal University of São Carlos
for data organization.
Data extraction
The reviewers underwent calibration exercises and worked in pairs to independently
extract data from the included studies according to the recommendations of the Cochrane
Handbook for Systematic Reviews of Interventions[20]. Disagreements were resolved by discussion or, if necessary, with third party adjudication.
Reviewers collected the following data using a pretested data extraction form: study
design, participants, interventions, comparators, assessed outcomes, and relevant
statistical data.
Risk of bias assessment
Two authors of this review independently assessed the risk of bias for each study
using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions[20]. Disagreements were resolved by discussion or by consultation with another review
author. We assessed the risk of bias according to the following domains.
-
Random sequence generation.
-
Allocation concealment.
-
Blinding of the participants and personnel.
-
Blinding of the outcome assessment.
-
Incomplete outcome data.
-
Selective outcome reporting.
-
Other biases.
We graded the risk of bias for each domain as high, low, or unclear and provided information
from the study report, together with justification for our judgment, in the “Risk
of bias” tables. For incomplete outcome data in individual studies, we stipulated
a low risk of bias for a loss to follow-up of less than 10% and a difference of less
than 5% in missing data between the intervention/exposure and control groups.
Certainty of evidence
We summarized the evidence and assessed its certainty separately for bodies of evidence
from RCT and non-RCT studies. We used the Grading of Recommendations Assessment, Development
and Evaluation (GRADE) methodology to rate the certainty of the evidence for each
outcome as high, moderate, low, or very low. In the GRADE approach, RCTs begin with
high certainty, and non-RCT studies begin with moderate certainty. Detailed GRADE
guidelines were used to assess the overall risk of bias, imprecision, inconsistency,
indirectness, and publication bias and to summarize the results in an evidence profile
([Table 1])[21].
Table 1
Characteristics of the included studies.
|
Author/Year
|
Study design
|
N/Age
|
Stroke time (days)
|
HRV variables
|
|
Nozoe et al., 2018[15]
|
Non-RCT
|
N = 21/71 y*
|
1 -10
|
LF, lnHF and LF/HF
|
|
Chen et al., 2019[23]
|
RCT
|
N = 72/65 ( 13.5 y
|
1 -10
|
SDNN, LF, HF and LF/HF
|
|
Beer et al., 2018[24]
|
Non-RCT
|
N = 19/63 y*
|
Post-acute
|
SDNN and RMSSD
|
|
Katz-Leurer & Shochina, 2007[22]
|
RCT
|
N = 64/62 ( 8.5 y
|
15
|
LF and HF
|
RCT: randomized clinical trial; HRV: heart rate variability; LF: low frequency; InHF:
natural logarithm of HF power; LF/HF: low to high frequency ratio; SDNN: standard
deviation of normal R-R intervals; RMSSD: root-mean-square difference of successive
normal R-R intervals; *The authors did not report the standard deviation.
We planned to assess publication bias through the visual inspection of funnel plots
for each outcome for which we identified 10 or more eligible studies; however, we
were not able to do so because there were an insufficient number of studies to conduct
this assessment.
Data synthesis and statistical analysis
It was not possible to perform a meta-analysis due to the non-homogeneity of the interventions.
The effects of the interventions, risk of bias, and quality of evidence for each study
are reported.
RESULTS
We identified a total of 88 studies through database searches (see [Figure 1] for the search results). After screening the titles and then the abstracts, we obtained
full-text articles for the 22 studies that were potentially eligible for inclusion
in the review. We excluded 18 studies because they were considered one of the following
types of articles: case report, case series, self-controlled study, review, or a study
that was not relevant. The remaining two RCTs[22],[23] and two non-RCTs[15],[24] were included in this review.
Figure 1 Flow chart of the search results.
Characteristics of the participants and groups
All participants in the included studies were diagnosed with ischemic stroke. The
total sample size was 172 individuals, and the average age was 65 years; they were
divided into groups, with the size of each group ranging from seven to 36 individuals.
In one study[15], there was no description of the difference between the intervention and control
groups since all of the participants received interventions; the participants were
instead divided according to stroke severity, as assessed by the National Institutes
of Health Stroke Scale (NIHSS). The other three studies[22]-[24] divided the individuals into intervention and control groups. [Beer et al. (2018]) described the control group as healthy individuals. Two studies[22],[24] included only individuals with one stroke, two evaluated patients within 1 to 10
days of an ischemic stroke[15],[23], one study evaluated individuals’ post-acute stroke[24], and another study evaluated individuals at 15 days after a stroke[22]. The characteristics of the included studies are shown in [Table 1].
All studies evaluated individuals based on the analysis of linear heart rate variables,
as shown in [Table 2].
Table 2
Variables evaluated in the four included studies.
|
HRV
|
System evaluated
|
|
SDNN
|
Sympathetic and parasympathetic activity
|
|
RMSSD
|
Parasympathetic activity
|
|
LF
|
Sympathetic and parasympathetic activity
|
|
HF
|
Parasympathetic activity
|
|
InHF
|
Parasympathetic activity
|
|
LF/HF
|
Sympathetic and parasympathetic balance
|
HRV: heart rate variability; SDNN: standard deviation of normal R-R intervals; RMSSD:
root-mean-square difference of successive normal R-R intervals; LF: low frequency;
HF: high frequency; InHF: natural logarithm of HF power; LF/HF: low to high frequency
ratio.
Evaluations and interventions
The interventions reported by the studies were early mobilization[15], low-intensity activity associated with meditation[23], cycle ergometer and cognitive activities[24], and protocol mobilization with a cycle ergometer, which were determined by exercise
resistance tests individually (cycle ergometer, walking test, and going up and down
stairs)[22]. All individuals in the control group performed activities such as conventional
physical therapy.
In the study by Nozoe et al. (2018)[15], the variables LH, InHF, and LF/HF ratio were evaluated by a cardiac monitor, and
in the analysis, the complement Lab Chart Pro HRV (ADInstruments Pty Ltd, Castle Hill,
Australia) was used. In the intervention protocol, the participants performed an early
mobilization in the sitting position; the evaluation comprised 5 minutes in the supine
position (rest), followed by five minutes in the sitting position. The patients were
reevaluated three months after the stroke.
In the study by Chen et al. (2019)[23], the variables SDNN, LF, HF, and LF/HF ratio were evaluated during the execution
of Chan-Chuang qigong, known as traditional Chinese medicine therapy, which promotes
body-mind interaction and relaxation. The individuals performed the technique for
15 minutes each day for 10 days; the assessment took five minutes and was performed
using a portable HRV analyzer (8Z11, Enjoy Research Inc., Taiwan), the Chinese version
of the Short Form-12 (SF-12) to assess quality of life, and the Hospital Anxiety and
Depression Scale (HADS) to assess negative emotions.
In the study by Beer et al. (2018)[24], individuals underwent a protocol in which they were first evaluated at rest for
10 minutes, and then they were evaluated during a handgrip activity that lasted two
minutes accompanied by controlled breathing (two minutes - six cycles in one minute).
Afterwards, they performed cognitive activity (serial 3’s subtractions) and finally
mobilization with a cycle ergometer in combination with a cognitive exercise. Cognitive
capacity was assessed using the Montreal Cognitive Assessment Scale (MoCA), and the
Barthel index was used to assess functional capacity. The variables SDNN and RMSSD
were measured by the Polar Advanced Heart Rate Monitor (RS800CX).
All included studies performed evaluations of linear heart rate variables; however,
studies did not present heterogeneity among the groups, interventions, or evaluations.
Only one study, of the four included, did not show significant results in relation
to the variables evaluated. All studies excluded individuals who had heart disease.
Evaluation of the effectiveness and safety of the included studies
The evaluation of the effectiveness and safety of the included studies are displayed
in the [Table 3].
Table 3
Interventions, results, and GRADE of the included studies.
|
Author (year)
|
Interventions
|
Results
|
GRADE
|
|
Nozoe et al., 2018[15]
|
Early mobilization in the ND group (n = 7) and early mobilization in the non-ND group
(n = 14)
|
LF/HF is higher in patients with ND during early mobilization; suggests increase in
sympathetic activity
|
⊕ Very Low
|
|
Chen et al., 2019[23]
|
Mind-body interactive exercise - Chan-Chuang qigong practice (15 minutes/day) (n =
36) and control group (n = 36)
|
LF/HF ratio significantly influenced the physical component of quality of life; suggests
cardiac autonomic balance after intervention
|
⊕⊕ Low
|
|
Beer et al., 2018[24]
|
Static and dynamic exercises; dual task; breathing exercise for 2 minutes.
|
Resting SDNN was significantly lower among post-stroke patients compared with healthy
individuals; less adaptive cardiac autonomic control during different activities
|
⊕ Very Low
|
|
Katz-Leurer & Shochina, 2007[22]
|
Conventional physical therapy
|
No alterations in HRV
|
⊕⊕ Low
|
ND: neurological deterioration classified by the NIHSS; LF: low frequency; HF: high
frequency; LF/HF: low to high frequency ratio; SDNN: standard deviation of normal
R-R intervals; HRV: heart rate variability. GRADE classification: • High-quality evidence:
Findings are consistent among at least 75% of the RCTs with a low risk of bias; data
are consistent, direct, and precise, and no publication biases are known or suspected.
Additional research is unlikely to change the estimate or our confidence in the results;
• Moderate-quality evidence: One of the domains is not met. Additional research is
likely to have an important impact on our confidence in the estimate of effect and
may change the estimate; • Low-quality evidence: Two of the domains are not met. Additional
research is very likely to have an important impact on our confidence in the estimate
of effect and is likely to change the estimate; • Very low-quality evidence: Three
of the domains are not met. We are very uncertain about the results.
In the study by Nozoe et al. (2018)[15], there were no significant differences in the InHF between the intervention and
rest values for the non-neurological deterioration (ND) group = 4.0 (3.2, 5.2) or
ND group = 4.7 (4, 5); P = 0.74; the LF/HF ratios were as follows: non-ND group = 1.9 (0.5, 3.2), ND group
= 1.0 (0.8, 3.3); P = 0.91. During mobilization, there were no significant differences in InHF between
the non-ND group = 4.9 (3.3, 5.9) and ND group = 4.6 (3.9, 4.9); P = 0.74. However, the LF/HF ratio was significantly higher in the ND group = 1.7 (SD
0.9, 2.6) than in the non-ND group = 0.6 (0.4, 1.5); P = 0.03. The authors did not report any adverse effects after intervention.
In the study by Chen et al. (2019)[23], the LF/HF ratio was higher in the intervention group after early mobilization regarding
the physical component of the quality of life (QOL) scale (SF-12) than in the control
group (P = 0.02). The authors did not report the effect sizes or confidence intervals of the
data, and any adverse effects were observed after intervention.
The study by Beer et al. (2018)[24] showed less adaptive cardiac autonomic control during different activities. The
values described for the groups were as follows: post stroke RR = 728.7 ± 110.1 ms;
healthy individuals RR = 847.6 ± 120.6 ms, with P = 0.002; post-stroke SDNN = 32.5 ± 26.9 ms, healthy individuals SDNN = 48.7 ± 17.9
ms, with P = 0.01. The authors did not report any adverse effects after intervention.
In the study by Katz-Leurer and Shochina (2007)[22], no significant interaction effects on HRV were observed between exercises during
physical therapy. The values indicated for the variables were as follows: treatment
group LF = 1248 ± 1684 Hz, control group LF = 1238 ± 1728 Hz, with P = 0.93; treatment group HF = 378 ± 638 Hz, control group HF = 667 ± 150 Hz, with
a P = 0.33. The authors did not report any adverse effects after intervention.
Risk of bias interpretation
All included articles were analyzed for risk of bias, as shown in [Table 4].
Table 4
Risk of bias classification.
|
Risk of bias
|
High Risk
|
Low Risk
|
Uncertain Risk
|
|
Random sequence generation
|
Beer et al., 2018[24]; Nozoe et al., 2018[15]
|
Chen et al., 2019[23]; Katz-Leurer; Shochina, 2007[22]
|
None
|
|
Allocation concealment
|
Beer et al., 2018[24]; Nozoe et al., 2018[15]
|
Chen et al., 2019[23]
|
Katz-Leurer; Shochina, 2007[22]
|
|
Blinding of the participants
|
Chen et al., 2019[23]
|
None
|
Katz-Leurer; Shochina, 2007[22] Beer et al., 2018[24]; Nozoe et al., 2018[15]
|
|
Blinding of the outcome assessment
|
Chen et al., 2019[23]
|
None
|
Katz-Leurer; Shochina, 2007[22] Beer et al., 2018[24]
; Nozoe et al., 2018[15]
|
|
Incomplete outcome data
|
Beer et al., 2018[24]; Nozoe et al., 2018[15]
|
Chen et al., 2019[23]
|
Katz-Leurer; Shochina, 2007[22]
|
|
Selective outcome reporting
|
None
|
Chen et al., 2019[23]
|
Katz-Leurer; Shochina, 2007[22] Beer et al., 2018[24]; Nozoe et al., 2018[15]
|
[Figure 2] shows a graphical analysis of the risk of bias.
Figure 2 Graphical analysis of the risk of bias in the included studies.
DISCUSSION
This systematic literature review study comprised four articles from clinical trials
that aimed to assess HRV using different methodologies, describing sympathovagal activity
after specific rehabilitation protocols in patients after ischemic stroke.
Of the four studies included, two[22],[24] used the cycle ergometer for the main rehabilitation program. Only the study by
Beer et al. (2018)[24] showed a significant reduction in the RR and SDNN variables among post-stroke individuals
compared to healthy individuals at rest, which indicates a state of sympathetic hyperactivity
in the subacute phase after the stroke. In this study, patients did not show a normal
increase in sympathetic activity in response to the test conditions, mainly due to
a hypersympathetic state at rest. During the subacute phase, according to this study
and other studies[25],[26], there is apparently a significant physiological change in the ability of the autonomic
nervous system to respond adequately to the demands imposed by rehabilitation practices,
so only large demands yield expected sympathetic responses[27]. The results indicate a need for rehabilitation focused on improving autonomic cardiac
control.
In the study by Nozoe et al. (2018)[15], patients were classified as having or not neurological deterioration (ND) using
the NIHSS score (severity scale used in the acute phase of stroke). These individuals
were evaluated during hospitalization and underwent an intervention involving early
mobilization with posture changes. The LF/HF ratio showed a significant increase in
the ND group (a higher NIHSS score) from before to after the intervention. Since the
LF/HF ratio seems to reflect sympathetic performance, according to the authors, it
is likely that an increase in sympathetic activity during mobilization is associated
with neurological deterioration in acute stroke patients. Xiong et al. (2018)[28] reported that autonomic dysfunction is one of the predictors of worse functional
outcomes in patients in the acute phase of stroke, which can confirm the possible
occurrence of increased sympathetic performance in patients with a worse NIHSS classification.
Chen et al. (2019)[23] introduced a mind-body interactive exercise (Chan-Chuang qigong practice) as an
intervention for hospitalized patients after stroke to increase cardiac parasympathetic
tone mainly because the technique has relaxing effects. They concluded that the LF/HF
ratio regarding the physical component of the quality of life (QOL) scale (SF-12)
was higher in the intervention group after mobilization than in the control group.
Therefore, during the hospital stay, the sympathovagal balance influenced the physical
aspect of the QOL of individuals with subacute stroke. Thus, improved HRV in stroke
patients after a specific rehabilitation protocol can lead to the recovery of physical
functions and improve their quality of life.
In the study by Katz-Leurer and Shochina (2007)[22], an individualized training protocol was used, and they did not find significant
differences in HRV. Despite this result, a significant improvement was found in the
functional parameters of post-stroke individuals, such as climbing stairs, and physical
training allowed patients to significantly increase their workload. As described by
other authors, autonomic impairment after stroke leads to low aerobic capacity[27]. Thus, the importance of early mobilization, rehabilitation, and physical-functional
training in post-stroke patients is reiterated.
The authors reported sympathetic-vagal alterations in poststroke patients when subjected
to physical activities. Thus, from this systematic review, it can be stated that significant
autonomic modulation occurs in these individuals. Despite the methodological divergence
found in the articles, only one article reported no changes in HRV between the groups
evaluated[22], which established an assessment in the frequency domain. In the study by Beer et
al. (2018)[24], variables in the time domain were included, whereas assessments in both domains
(time and frequency) were included in other studies, which demonstrated significant
changes in the HRV linear variables after stroke rehabilitation.
Studies on HRV demonstrate the need for flexibility in autonomic activity for individuals
to maintain a good quality of life, as impaired adaptation can cause autonomic dysfunctions,
cardiovascular deterioration, and increased morbidity and mortality rates in patients
after stroke[28]. The four articles selected for the review show the need for specific therapies,
early mobilizations, and physical activity protocols in the modulation of HRV. This
conclusion points to the importance of maintaining muscle function, strength, and
activity for cardiovascular benefits, which has been widely studied for methods including
cardiac rehabilitation[28]
-
[30].
This study has limitations, such as heterogeneity in the selected individuals and
the analyzed outcomes; because only a few studies were selected, it was impossible
to perform a meta-analysis. However, this is the first systematic review addressing
this topic, with the possibility of elucidating the main autonomic repercussions observed
in post-stroke patients undergoing rehabilitation procedures.
In conclusion, the quality of the evidence from the selected clinical trials was either
low or very low; therefore, there are no definitive conclusions about the main autonomic
repercussions observed in post-stroke patients undergoing rehabilitation, although
all interventions are safe for these patients. The applicability of these results
can be compromised since most of the results described in this review were obtained
from clinical trials with methodological differences. This review highlights the need
to conduct well-designed tests in this field. Future trials should be properly designed
and should include standardized measures. It is suggested that RCTs address a heterogeneous
population and include measures in the time and frequency domains, in addition to
a nonlinear analysis of HR, to establish parameters of sympathetic-vagal behavior
during rehabilitation protocols after stroke.
