Development of a prediction model for ascent and descent staircase independence during the sub-acute rehabilitation phase in individuals post-stroke

Abstract

BACKGROUND:

Nearly half of individuals post-stroke are dependent in their daily activities. The ability to ascend and descend stairs is an important component of independence in the community.

OBJECTIVE:

To predict the future ability of post-stroke individuals at the beginning of the sub-acute rehabilitation phase to achieve ascending and descending staircase independence.

METHODS:

36 participants were recruited for the study. Outcome measures: independence in ascending and descending stairs up to the end of the sub-acute rehabilitation phase. Predictive measures included the knee extensors Muscles Strength (MS), the seated Modified Functional Reach (MFR) test, and Heart Rate Variability (HRV) measures. Logistic and Cox regression were used.

RESULTS:

Twenty-four participants (66.7%) completed the sub-acute rehabilitation phase being independent in ascending and descending stairs. MFR was the best predictor (R² = 0.18), and with MS the best predictors for the time (days) to achieve this goal during the sub-acute rehabilitation. HRV measure was found to be the main predictor of the staircase ascent model (R² = 0.32), and MFR the best predictor for the descent model (R² = 0.24).

CONCLUSIONS:

Balance performance is the main predictor of independence in ascending and descending stairs. Improving this component during the sub-acute rehabilitation phase might be reflected in achieving staircase independence.

Keywords

Stroke stairs ascending and descending heart rate variability muscles strength balance performance

1 Introduction

Stroke is the leading cause of long-term disability worldwide (Moftah & Jadavji, 2019). Nearly half of individuals post-stroke present motor impairments that lead to dependence in their daily activities (Bonita, Solomon & Broad, 1997), and more than half report restrictions in participation (Mayo, Wood-Dauphinee, Côte, Durcan & Carlton, 2002). Walking and balance restriction, as well as the inability to ascend and descend stairs, lead to limitations in social life (Hinman, O’Connell, Arnold, Chandler & Flores, 2017; Olney, Griffin & McBride, 1994; Robertson et al., 2009; Wilkinson et al., 1997). Ascending and descending stairs has been found to be the best predictor of activity level in the elderly post-stroke (Alzahrani, Dean & Ada, 2009).

The ability to ascend and descend stairs requires sufficient muscle strength, coordination, stability and aerobic endurance (Novak & Brouwer, 2012; Park, Kim & Kim, 2015). These components are frequently impaired after stroke (Alzahrani et al., 2009), and often results in participation restriction; deficits in strength and motor coordination related to walking and activity restrictions (Menezes et al., 2018), balance impairments associated with low ambulatory activity and low cardiovascular fitness which influence the activities of daily living (Michael, Allen & Macko, 2005). Ascending stairs requires simultaneous movement on the horizontal and vertical planes in order to maintain balance (Park et al., 2015). Furthermore, the effort increases with the need to cope with gravity and maintain stability outside the base of support (Lee & Seo, 2014).

In cross-sectional studies with individuals post-stroke in the chronic phase, associations between balance performance (assessed by the Berg Balance Scale [BBS]) and the ability to ascend stairs was noted (Fujita et al., 2018). Furthermore, in a cross-sectional study among patients at the acute, sub-acute and chronic phases post-stroke (4–347 days post-stroke) balance performance and lower limb muscle strength explain 84.6%of the variance of stair-climbing ability (Bohannon & Walsh, 1991). Cross-sectional studies also found that walking speed, walking distance, and the strength of the straightening muscles of the knee explained about 50%of the variance in ascending and descending stairs (Polese, Scianni & Teixeira-Salmela, 2015). It has also been found that components of strength and coordination of the muscles of the lower limbs are related to the time required to ascend and descend stairs (Menezes et al., 2018). The stronger the muscles of the lower limbs, the lower the spasticity, and the longer the standing time on the weaker leg, the shorter the time required for ascending and descending the stairs, making performance more effective (Bonnyaud, Zory, Pradon, Vuillerme & Roche, 2013).

Regarding the aerobic requirement, an increase in the energy consumption around the knee and ankle muscles observed while ascending and descending stairs (Novak et al., 2012) and the higher the functional impairments are, the higher the energy cost in fulfilling this task (Polese, Ribeiro-Samora, Lana, Rodrigues-De-Paula & Teixeira-Salmela, 2017). Energy consumption is higher than walking, with an increase in VO2 Max (Maximum oxygen uptake) consumption along with an adjustment of heart rate and respiration to the change in energy demand during ascending and descending stairs (Modai, Sharon, Bar-Haim & Hutzler, 2015). Beyond that, physiological measures, such as ventilation, oxygen uptake and heart rate values were found to be significantly higher in step to ascending type as compared to the step over pattern. Also the caloric expenditure in the step to ascending type was slightly higher (Aziz & The, 2005). Meaning that the step over ascending and descending type is the effective and typical pattern.

Aerobic fitness level was found to be associated with heart rate variability (HRV); the greater the aerobic capacity, the better and more variable the HRV in an aging population (De Meersman, 1993). In addition, the HRV index was found to predict endurance in walking in patients post-stroke (Shapira-Vadler, Treger & Katz-Leurer, 2015). HRV is a marker for autonomic nervous system activity and is defined as the variation in time interval between R-R intervals. Individuals post-stroke present lower HRV parameters which are associated with an increased sympathetic tone (Raimundo et al., 2013). It was found that HRV measures during the initial phase of rehabilitation predict (r = 0.6) walking distance during the 6 Minute Walk Test (6MWT), tested one month later (Shapira-Valder et al., 2015).

The current study aims to predict at the beginning of the sub-acute rehabilitation phase the ability of a patient post-stroke to independently and efficiently ascend and descend a staircase at the end of this phase. As with the above cross-sectional studies, associations between different functions and abilities in ascending and descending stairs were present at different stages post-stroke. In the current follow-up study the aim was to assess which of these characteristics can predict staircase independence.

2 Methods

2.1 Study population

Participants between the ages of 18–75 years were recruited. Included were patients diagnosed with an ischemic or hemorrhagic first stroke based on medical records, physical examination and imaging findings, and patients with sufficient communication and cognitive abilities to understand instructions. Excluded were patients with post-traumatic head injury, those who were in a confused state and uncooperative, patients with complex cardiovascular diseases that prevented intense activity, and patients with the ability to ascend and descend stairs independently.

The study protocol was approved by the Helsinki Committee of ‘Sheba’ Tel-Hashomer medical center and the Ethics Committee of Tel-Aviv University. Participants were recruited by the first author (G.R) and all signed an informed consent form prior to participation.

2.2 Research tools

Outcome measures: 1) Success in ascending and descending a 13-step staircase independently with or without device, with a score of 6-7/7 on the Functional Independence Measure (FIM), which was found a valid assessment tool for individuals post-stroke (Concurrent Validity) (Hsueh, Lin, Jeng & Hsieh, 2002). Achieving independence categorized as staircase independence sub-group, otherwise as dependence sub-group. 2) Time (in days) from admission to the sub-acute rehabilitation department and up to when achieving independence, or up to discharge from the department without achieving this goal; 3) Ascending and descending stairs model (step over or step to); 4) Time (in seconds) to ascend and descend.

To assess the predictive measures, HRV was monitored by a Polar RS800CX watch (Polar Electro OY watch, Kempele, Finland) and calculated by the time indices: a) R-R intervals; b) the standard deviation of normal to normal intervals (SDNN) which represents the overall HRV; c) The root mean square of the successive differences (RMSSD) which estimates the short-term variability of the R-R intervals. All were recorded in units of millisecond (ms) (Task Force of the European Society of Cardiology, 1996). The data was received from a transmitter placed on the chest and then transmitted to a computer which was then analyzed by software designed for Polar (Polar Protrainer 5 Inc.). In validating the polar S810 watch (Polar Electro OY, Kempele, Finland), validity criterion showed a high correlation between the information obtained from the Polar and the information obtained from an electrocardiogram. In short-term five-minute measurements among healthy young subjects the reliability of repeat measurement was found to be high (the Intraclass Correlation Coefficient [ICC] index ranged from 0.85–0.99) (Nunan et al., 2008; Farah et al., 2016).

Muscle Strength (MS) of the knee extensors in the weaker leg, estimated by Motricity Index (MI) which was found valid and reliable in assessing muscle strength in individuals post-stroke, and correlated highly with objective quantitative measures obtained by Hand-Held Dynamometry (HHD) (Cameron & Bohannon, 2000; Fayazi, Dehkordi, Dadgoo & Salehi, 2012). The results of this test were based on the Medical Research Council (MRC) index divided into two sub-groups; 4-5/5 versus (vs.) 0–3/5. These categories were chosen based on the assumption that muscle strength of 4-5/5 is needed for ascending stairs independently.

Modified Functional Reach (MFR) estimated the maximum forward bending distance while sitting. This test evaluates the ability to maintain stability outside the base of support. The test was found valid for individuals post-stroke in the sub-acute phase vs. the balance master, stroke activity scale and FIM (Katz-Leurer, Fisher, Neeb, Schwartz & Carmeli, 2009).

The anthropometric and demographic characteristics including age, gender, weight and height were taken from the medical records as were such stroke characteristics as type and side weakness. Data on exercise, smoking and health condition before the event were also collected.

2.3 Study design

The current study is a follow-up performed at the Brain Injury Rehabilitation department at ‘Sheba’ Tel-Hashomer medical center.

All inpatients were examined in the first 48 hours of admission by the physiotherapy team. Patients who met the inclusion criteria signed the informed consent form and were examined over the next four days by the first author (G.R) as part of the follow-up study. HRV (including RR, SDNN and RMSSD values) was measured at rest in the sitting position for five minutes at rest, during five minutes of cycling and for five minutes in recovery. Additionally, MS of the weaker leg was measured according to MI for each participant in a sitting position in a wheelchair, and finally, the MFR test was performed with the less affected arm in the same sitting position.

As part of the rehabilitation routine all the participants received daily physiotherapy, occupational therapy, speech therapy and psychological counseling. Physiotherapy focused on improving functional abilities such as transitions, walking and stairs. Upon achieving stair independence, the study participants were re-tested by the first author (G.R). HRV was measured for five minutes at rest, while ascending and descending stairs, and for five minutes during recovery.

2.4 Statistical analysis

Sample size calculation was based on the study by Shapira-Valder et al. (2015), which noted a positive moderately strong predictive association between HRV and 6MWT results. With a type one error of 5%and a test power of 80%the minimum sample size needed was 39 participants in total. In addition, a preliminary survey in 2017, found a ratio of 2 : 1 in favor of those who achieved the ability to ascend and descend stairs compared to those who did not. Therefor for this study 26 participants were calculated for achieving independence and 13 not. Recruitment was halted after 36 participants due to the outbreak of the coronavirus disease (COVID-19) with new guidelines regarding the conduct of clinical trials.

Description of characteristics of all participants and of the independence/dependence sub-groups were performed using frequency and relative frequency for nominal variables, median and range for ordinal variables, and mean and standard deviation for ratio variables. Assumption of a normal distribution was tested for all ratio variables using the Kolmogorov-Smirnov test. If the assumption was violated, the variables were indicated by median and range.

Differences in the predictive measures at admission between the sub-groups are shown using x² test, t-test, or Mann-Whitney U test, dependent on normal assumption violation or not, as well as the associations, Spearman vs. Pearson, respectively. In an attempt to identify an optimal intersection point for predicting ascending and descending stairs according to the predictive measures, a Receiver Operating Characteristic (ROC) diagram was used. To examine the contribution of a combination of predictive measures on outcome measures, multivariate, logistic and Cox regression were used. To determine the best predictors, the stepwise regression included thresholds of p = 0.05 for entry and 0.10 for removal. A test was defined as significant for p-value < 0.05, and data analysis was performed using SPSS version 25 software (SPSS Inc., USA).

3 Results

3.1 Participant’s characteristics

Twenty-four out of 36 participants achieved staircase independence according to FIM (FIM 6-7/7) during the sub-acute rehabilitation phase. Table 1 presents the background and stroke characteristics of the participants, in the entire sample and according to independence/dependence sub-groups.

Table 1
Participants characteristics of the entire sample and by independence/dependence sub-groups

Characteristics Sub-groups

All N = 36 Independence N = 24 Dependence N = 12 p-value

Socio-demographic

Age (years) 56.36±12.01 59.5 [29.0–73.0] 62.0 [23.0–72.0] 0.57

Gender (men) 24 (66.7) 15 (62.5) 9 (75.0) 0.45

Health habits

Body mass index (kg/meter²) 27.36±5.22 25.7 [21.4–40.6] 25.8 [21.4–46.5] 0.97

Exercise (days a week) 1.22±1.79 1 [0–7] 0 [0–2] 0.02

Smoking (no) 33 (91.7) 2 (8.3) 1 (8.3) 1.00

Comorbidities

Hypertension 27 (75.0) 17 (70.8) 10 (83.3) 0.41

Hyperlipidemia 17 (47. 2) 12 (50.0) 5 (41.7) 0.64

Diabetes mellitus 11 (30.6) 5 (20.8) 6 (50.0) 0.07

Stroke

Hemorrhagic 20 (55.6) 15 (62.5) 5 (41.6) 0.48

Side weakness (left) 19 (54.3) 12 (52.2) 7 (58.3) 0.73

Characteristics		Sub-groups
Socio-demographic
Age (years)	56.36±12.01	59.5 [29.0–73.0]	62.0 [23.0–72.0]	0.57
Gender (men)	24 (66.7)	15 (62.5)	9 (75.0)	0.45
Health habits
Body mass index (kg/meter²)	27.36±5.22	25.7 [21.4–40.6]	25.8 [21.4–46.5]	0.97
Exercise (days a week)	1.22±1.79	1 [0–7]	0 [0–2]	0.02
Smoking (no)	33 (91.7)	2 (8.3)	1 (8.3)	1.00
Comorbidities
Hypertension	27 (75.0)	17 (70.8)	10 (83.3)	0.41
Hyperlipidemia	17 (47. 2)	12 (50.0)	5 (41.7)	0.64
Diabetes mellitus	11 (30.6)	5 (20.8)	6 (50.0)	0.07
Stroke
Hemorrhagic	20 (55.6)	15 (62.5)	5 (41.6)	0.48
Side weakness (left)	19 (54.3)	12 (52.2)	7 (58.3)	0.73

Values in table are number (percentage), median [min-max] and average±standard deviation. p-value based on Chi-Square test, Mann-Whitney U test, or t-test respectively.

The median age of the participants was 61.5, with ages ranging from 29–73 years. Twenty-one of the participants (55.6%) had a Body Mass Index (BMI) value higher than the threshold value of 25, which is defined as overweight by the World Health Organization (WHO). (Retrieved from https://www.who.int/gho/ncd/risk_factors/bmi_text/en/). Thirty-one of the participants (86.1%) had one or more pre-stroke co-morbidities, with hypertension (75% of the sample) being the most prevalent. No significant differences in background characteristics were found between the sub-groups, except for pre-stroke exercise frequency; 50% of those who achieved independence did at least one exercise session per week, while 50% or more of those who did not achieve performed no exercise at all (Table 1).

Among the 24 participants of the independence sub-group, six used an Ankle Foot Orthosis (AFO) and one used a dictus. Except for one participant, everyone used the railing. The median time for achieving staircase independence was 73 days (Table 2).

Table 2

Ascending and descending stairs

Characteristics	N = 24
Days in rehabilitation for achieving independence	73 [18–192]
Ascending model
Step over	16 (66.7)
Step to	8 (33.3)
Descending model
Step over	12 (50.0)
Step to	12 (50.0)
Ascending time (seconds)	15.0 [6.0–49.0]
Descending time (seconds)	18.5 [6.0–64.0]

Values in table are number (percentage) and median [min-max].

3.2 Independence in ascending and descending stairs –univariate analysis

The aim of the current study was to predict the future ability of post-stroke individuals at the beginning of the sub-acute rehabilitation phase to achieve ascending and descending staircase independence. Tables 3 4 present these data using univariate analysis.

Table 3
Functional and autonomic characteristics of the entire sample and by independence / dependence sub-groups

Characteristics Sub-groups

All N = 36 Independence N = 24 Dependence N = 12 p-value

Functional

MS (4-5/5) (affected leg) 10 (27.8) 7 (29.2) 3 (25.0) 0.79

MFR (cm) (less affected hand) 34.5±9. 9 36.50±8.47 30.50±11.60 0.04

HRV measures at rest

R-R (ms) 663.0 [414.0–1069.0] 661.5 [414.0–1062.0] 707.0 [419.0–1069.0] 0.97

SDNN (ms) 19.8 [4.7–53.0] 19.6 [4.7–37.0] 21.5 [9.3–53.0] 0.26

RMSSD (ms) 12.1 [1.8–62.0] 7.6 [1.8–37.0] 13.6 [3.9–62.0] 0.22

HRV measures differences between rest and cycling

R-R (ms) delta –17.0 [(–147.0)–244.0] –19.0 [(–147.0)–244.0]^* –12.5 [(–43.0)–180.0] 0.16

SDNN (ms) delta –1.6 [(–25.1)–25.0] –1.0 [(–25.1)–25.0]^* –2.6 [(–21.1)–16.0] 0.32

RMSSD (ms) delta –0.4 [(–13.4)–52.0] –0.2 [(–13.4)–52.0] –1.4 [(–9.9)–8] 0.92

Characteristics		Sub-groups
Functional
MS (4-5/5) (affected leg)	10 (27.8)	7 (29.2)	3 (25.0)	0.79
MFR (cm) (less affected hand)	34.5±9. 9	36.50±8.47	30.50±11.60	0.04
HRV measures at rest
R-R (ms)	663.0 [414.0–1069.0]	661.5 [414.0–1062.0]	707.0 [419.0–1069.0]	0.97
SDNN (ms)	19.8 [4.7–53.0]	19.6 [4.7–37.0]	21.5 [9.3–53.0]	0.26
RMSSD (ms)	12.1 [1.8–62.0]	7.6 [1.8–37.0]	13.6 [3.9–62.0]	0.22
HRV measures differences between rest and cycling
R-R (ms) delta	–17.0 [(–147.0)–244.0]	–19.0 [(–147.0)–244.0]^*	–12.5 [(–43.0)–180.0]	0.16
SDNN (ms) delta	–1.6 [(–25.1)–25.0]	–1.0 [(–25.1)–25.0]^*	–2.6 [(–21.1)–16.0]	0.32
RMSSD (ms) delta	–0.4 [(–13.4)–52.0]	–0.2 [(–13.4)–52.0]	–1.4 [(–9.9)–8]	0.92

Values in table are number (percentage) median [min-max] and mean±standard deviation. p-value base on Mann-Whitney U test or t-test respectively. ^*p-value < 0.05 base on Wilcoxon test. MS- Muscles Strength, MFR- Modified Functional Reach, HRV- heart rate variability, SDNN- standard deviation of normal to normal intervals, RMSSD- The root mean square of the successive differences.

Table 4

Functional and autonomic characteristics at admission by model of ascending and descending stairs at follow-up (N = 24)

	Ascending			Descending
Characteristics	Step over N = 16	Step to N = 8	p-value	Step over N = 12	Step to N = 12	p-value
MS (0–5) affected leg	2.0 [0.0–5.0]	1.5 [0.0–3.0]	0.01	3.0 [0.0–5.0]	2.0 [0.0–4.0]	0.03
MFR (cm) less affected hand	40.0 [28.0–52.0]	33.5 [19.0–44.0]	0.02	41.0 [28.0–52.0]	33.5 [19.0–45.0]	0.02
SDNN (ms) delta rest vs. cycling	3.6 [(–3.3)–25.1]^*	–0.1 [(–25.0)–8.4]	0.01	4.0 [(–3.3)–12.1]	0.2 [(–25.0)–25.1]	0.04

Values in table are median [min-max]. p-value base on Mann-Whitney U test. ^*p-value < 0.05 base on Wilcoxon test. MS- Muscles Strength, MFR- Modified Functional Reach, SDNN- standard deviation of normal to normal intervals.

The average MFR was six centimeters (cm) farther among the independence sub-group (p = 0.04). No significant difference in MS distribution between sub-groups was noted. The median R-R intervals of all participants was 663.0 [414.0–1069.0], which represents a mean heart rate value of 90.5 beats per minute. No significant difference in HRV measures were noted between the sub-groups. A significant change was noted in the independence sub-group in R-R interval and SDNN values between rest and cycling upon admission to sub-acute rehabilitation. (Table 3).

Time (in days) from admission to the sub-acute rehabilitation and up to achieving staircase independence (independence sub-group) or upon discharge without achieving this goal (dependence sub-group) was found to be negatively and significantly associated with MFR (r_s = –0.39, p-value = 0.02). In addition, the time was significantly different between participants with MS 4-5/5 (median 46 [18–116] according to MI values) compared to participants with MS 0–3/5 (median 94 [28–208] according to MI values) (p-value = 0.02). No associations between time and HRV measures were noted. These data are not presented in the tables.

3.3 Stairs ascending and descending model - univariate analysis

Significant differences were noted in MS, MFR and SDNN (rest vs. cycling) between those who ascended or descended the stairs in the step over model as compared to the step to model. The first presented higher MS, farther MFR and a greater change in SDNN due to cycling (Table 4). In addition, only those who ascended or descended in the step over model presented a significant change in SDNN values between rest and cycling at admission to sub-acute rehabilitation (Table 4). No significant predictive associations were noted between predictor variables and time needed to ascend and descend the staircase. These data are not presented in the table.

3.4 Multivariate prediction analysis

Multivariate analyses were used for assessing the best predictor for each outcome: see results presented in Table 5. For each model only the variables that were found significantly predictors in the univariate analysis were entered. The MFR was entered to a logistic regression model for predicting staircase independence with a cutoff value of±32.5 cm. This value was the optimal value for the accuracy index as previously found using ROC analysis (Sensitivity –71%, specificity –87.5%, area under the curve 69%). The MFR explains 18% of staircase independence. The odds ratio for participants who achieved more than 32 cm for independence was 7.29 as compared to those who achieved 32 cm or less.

Table 5
Logistic models for best predictors of outcome measures

Outcome measure Model Characteristic Regression coefficient Standard error Odds ratio/Hazard ratio 95% confidence interval

Staircase independence Logistic MFR > 32 (cm) 1.99 0.80 7.29 1.50–35.21

Time (days) for independence Cox MFR > 32 (cm) –2.05 0.56 0.13 0.04–0.38

MS 4-5/5 –1.81 0.56 0.16 0.05–0.49

Ascending model Logistic SDNN delta > 1 (ms) 1.72 0.80 5.60 1.16–27.06

Descending model Logistic MFR > 38 (cm) 2.02 0.83 7.50 1.47–38.28

Outcome measure	Model	Characteristic	Regression coefficient	Standard error	Odds ratio/Hazard ratio	95% confidence interval
Staircase independence	Logistic	MFR > 32 (cm)	1.99	0.80	7.29	1.50–35.21
Time (days) for independence	Cox	MFR > 32 (cm)	–2.05	0.56	0.13	0.04–0.38
		MS 4-5/5	–1.81	0.56	0.16	0.05–0.49
Ascending model	Logistic	SDNN delta > 1 (ms)	1.72	0.80	5.60	1.16–27.06
Descending model	Logistic	MFR > 38 (cm)	2.02	0.83	7.50	1.47–38.28

MS- Muscles Strength, MFR- Modified Functional Reach, HRV- heart rate variability, SDNN- standard deviation of normal to normal intervals.

To predict time (in days) to achieve independence during sub-acute rehabilitation, MFR and MS were entered to a Cox regression model. A strong, negative and significant predictive association was found between MFR > 32 and time to independence (hazard ratio of 0.13). Similarly, a strong, negative and significant predictive association was found between MS 4-5/5 and time to independence (hazard ratio of 0.16) (Table 5). The survival curve of each predictor is shown in Fig. 1.

Fig. 1

Survival curve for time to independence or up to end of the sub-acute rehabilitation by Modified Functional Reach (MFR) and Muscles Strength (MS) sub-categories (the broken line represent 50% cumulative survival rate).

For the ascending model, the change in SDNN value between rest and cycling at admission to the sub-acute rehabilitation was the only variable remained in the model and explained 32% of this outcome variance. A decrease of at least one ms of SDNN showed maximal accuracy, as previously found by using ROC analysis, to predict ascending stairs in the step over model (Sensitivity –71%, specificity –87%, area under the curve 0.84). In participants who presented a change of one ms or more between rest and cycling, the odds ratio for the step over model was 5.6 as compared to those with a change of less than one ms.

For the descending model, a MFR measure of±38 cm was the only variable remained in the model and explains 24% of this outcome variance. The cutoff value of±38.0 cm was the optimal value for the accuracy index previously found by using ROC analysis (Sensitivity –69%, specificity –89%, area under the curve 72%). For participants who achieved more than 38 cm the odds ratio for the step over model was 7.5 as compared to those who achieved 38 cm or less.

4 Discussion

The aim of the current study was to predict the future ability of individuals after stroke to ascend and descend a 13-step staircase independently during the sub-acute rehabilitation phase. MFR, MS and HRV measures were assessed as potential predictors at admission to rehabilitation.

Twenty-four of the 36 participants achieved independence in the sub-acute rehabilitation phase. It was found that the MFR is the best predictor, and was also found to be the main predictor of the stairs descending model. MFR and MS were the best predictors for time in days to achieving independence. The autonomic characteristics did not predict independence, but the SDNN delta value between rest and cycling at admission to rehabilitation was found to be the main predictor of the stairs descending model.

4.1 Balance measure is the only predictor for staircase independence

This finding is consistent with previous studies. Bohannon et al. (1991), showed among a heterogeneous population post-stroke that balance ability in standing is strongly associated (r = 0.89) with staircase independence (Bohannon et al., 1991). In the Bohannon et al. study the MFR variance was 79% vs. a MFR variance of 18% in the current study. Such a dramatic difference; a strong association, vs. 18%, which is relatively weak, may be due to the fact that Bohannon et al. performed a cross-sectional study with a very heterogeneous group of patients (N = 20); the participants were tested between 4–347 days after the event, while in the current study a more homogeneous group at the sub-acute phase post-stroke, was assessed at the beginning of rehabilitation in order to predict ability at the end. However, although significant differences in study design and time post-stroke exist between the studies, balance performance in both is the main variable associated with stair function ability.

4.2 Functional measures are predictors for the time to functional independence

In a comprehensive review we did not find any follow up or retrospective studies which assessed the prediction of or the associations between functional and physical abilities and the time necessary for achieving staircase independence. Although in previous studies muscle strength and balance performance were found to be associated with stair independence (Menezes et al., 2018, Fujita et al., 2018, Bohannon & Walsh, 1991), it may simply be that individuals post-stoke who attend a rehabilitation department with better MS and balance performance, have a greater ability to achieve independence earlier.

4.3 SDNN measure is the primary predictor of the stairs ascending model

The autonomic characteristics predicted the ascending model among participants who achieved independence for which there are two possible explanations. The first is the physiological effort of ascending stairs and is supported by Shapira-Valder et al. (2015); SDNN at rest at admission to rehabilitation was the best predictor for the 6MWT results a month later (Shapira-Valder et al., 2015). Since ascending and descending stairs increases oxygen consumption (VO₂), heart-rate (HR), and metabolic equivalents (MET) relative to walking among individuals post-stroke (Polese et al., 2017), therefore the autonomic measure (HRV) is the best predictor for the ascending model.

A second explanation may be based on the association that exists between stair performance and motor ability and HRV and motor ability in patients post-stroke. It was previously found that autonomic impairment is associated with motor performance in patients post-stroke (Raphaely & Katz-Leurer, 2014), and that decreased HRV, and SDNN specifically, is associated with motor impairments in these patients (Bassi, Colivicchi, Santini & Caltagirone, 2007). Moreover, Graff et al. (2013) found a predictive association between HRV and functional ability according to the National Institutes Of Health Stroke Scale (NIHSS) and Modified Rankin Scale (MRS) one week after a stroke (Graff et al., 2013); Tang et al. (2015) found that the RMSSD value in the first week after stroke predicted MRS results three months later (Tang et al., 2015); Arad et al. (2002) found a medium to strong predictive association between HRV (SDNN, LF, HF) and functional abilities according to the FIM scale in admission and discharge from the hospital (Arad, Abboud, Radai & Adunsky, 2002).

4.4 Study limitations

Based on the primary sample size analysis our goal was to recruit 39 patients post-stroke, based on our preliminary assumption that by the end of the rehabilitation period, 26 will achieve staircase independence while 13 will not. Only 36 participants were finally enrolled due to the outbreak of the Corona virus disease (COVID-19). Moreover, based on a post hoc sample size calculation related to the addition of HRV measures to the ascending model, the sample size needed was 42 participants.

The participants in this study were younger than the post-stroke population in Israel. According to the Israel Neurological Organization, the average age for a stroke in Israel is 68.6±12.9 years among men and 73.0±13.6 among women (National Acute Stroke Israeli [NASIS], registry, 2013). In contrast, this study found that the mean age among men is 57.87±10.13 and among women, 53.33±15.14. The main reason is that this study was performed in a hospital with two neurologic rehabilitation departments, one in which this study was performed in and another for the geriatric population.

Unlike a relative incidence of about 40% of ischemic strokes in this study, the incidence of ischemic strokes is much higher (92.6%), and only 6.9% are cerebral hemorrhages (National Acute Stroke Israeli [NASIS], registry, 2013). The high incidence of cerebral hemorrhage in this study may be explained by selective referral according to the type of stroke to sub-acute rehabilitation. Despite the known data indicating higher rates of ischemic strokes (Shiber, Fontane & Adewale, 2010), the NASIS shows that cerebral hemorrhage causes more severe injury. In addition, about half of the individuals after cerebral hemorrhage were transferred from the acute departments to the rehabilitation department, compared to only 29% of the individuals after an ischemic stroke (National Acute Stroke Israeli [NASIS], registry, 2013).

Regarding the manipulation of the study, the question arises whether the cycling in the manner in which it was performed was a differential test. In a healthy population, it is known that during aerobic exercise the HR control system responds with an increase in HR and a decrease in HRV (Kaikkonen, Nummela & Rusko, 2007). However, it is possible that the motor impairment among the participants prevented them from increasing the pace and resistance during cycling, resulting in an satisfactory aerobic challenge, and therefore the necessity for cardiac autonomic adjustment was less. Similarly, regarding the outcome measures, it is possible that if a different, more complex, aerobic effort had been selected, other results would have obtained. It is possible that climbing a larger staircase would have increased energy consumption and thus the necessity for cardiac autonomic adjustment and adequate aerobic capacity.

4.5 Conclusion and recommendations

Neurological rehabilitation is complex and our ability as clinicians to identify those individuals after damage to the central nervous system with future functional independence ability is difficult and problematic. Determining clear characteristics, as found in this study: MS over 3/5 and MFR farther than 32 cm, can help assess the likelihood of functional independence and can help us estimate the time required for hospitalization (as found in this study as a median time of 73 days). These data do not indicate a dichotomy between the functional ability of independence / dependence in ascending and descending stairs, but they can help us assess rehabilitation expectations and advise families in matters relating to housing adjustment, employment and quality of life.

As found in this study, the balance component is the main predictive measure. Although there are many different rehabilitation models used around the world, estimating the MFR at the beginning of rehabilitation and dividing it into below or above 32.5 cm can predict staircase independence (sensitivity –71%, specificity –87.5%) at the end of rehabilitation. To validate this model, more research with a similar population in other places is needed. it is possible that the improvement of this component may be reflected in achievement of independently ascending and descending stairs, and therefore it is important to integrate it frequently and primarily in physiotherapy treatment, as part of sub-acute rehabilitation. Further studies are needed to confirm this hypothesis.

Conflict of interest

None to report.

References

Moftah,

, & Jadavji,

N. M.

(2019). Role of behavioral training in reducing functional impairments after stroke. Neural Regeneration Research, 14(9), 1507.

Bonita,

, Solomon,

, & Broad,

J. B.

(1997). Prevalence of stroke and stroke-related disability: estimates from the Auckland Stroke Studies. Stroke, 28(10), 1898–1902.

Mayo,

N. E.

, Wood-Dauphinee,

, Côte,

, Durcan,

, & Carlton,

(2002). Activity, participation, and quality of life 6 months poststroke. Archives of Physical Medicine and Rehabilitation, 83(8), 1035–1042.

Hinman,

M. R.

, O’Connell,

J. K.

, Arnold,

L. A.

, Chandler,

, & Flores,

(2017). Functional Predictors of Stair-Climbing Ability in Older Adults. MOJ Gerontol Ger, 1(5), 00025.

Olney,

S. J.

, Griffin,

M. P.

, & McBride,

I. D.

(1994). Temporal, kinematic, and kinetic variables related to gait speed in subjects with hemiplegia: a regression approach. Physical Therapy, 74(9), 872–885.

Robertson,

J. V.

, Pradon,

, Bensmail,

, Fermanian,

, Bussel,

, & Roche,

(2009). Relevance of botulinum toxin injection and nerve block of rectus femoris to kinematic and functional parameters of stiff knee gait in hemiplegic adults. Gait & Posture, 29(1), 108–112.

Wilkinson,

P. R.

, Wolfe,

C. D.

, Warburton,

F. G.

, Rudd,

A. G.

, Howard,

R. S.

, Ross-Russell,

R. W.

, & Beech,

R. R.

(1997). A long-term follow-up of stroke patients. Stroke, 28(3), 507–512.

Alzahrani,

M. A.

, Dean,

C. M.

, & Ada,

(2009). Ability to negotiate stairs predicts free-living physical activity in community-dwelling individuals with stroke: an observational study. Journal of Physiotherapy, 55(4), 277–281.

Novak,

A. C.

, & Brouwer,

(2012). Strength and aerobic requirements during stair ambulation in persons with chronic stroke and healthy adults. Archives of Physical Medicine and Rehabilitation, 93(4), 683–689.

10.

Menezes,

K. K.

, Nascimento,

L. R.

, Faria,

C. D.

, Avelino,

P. R.

, Scianni,

A. A.

, Polese,

J. C.

,... & Teixeira-Salmela,

L. F.

(2018). Deficits in motor coordination of the paretic lower limb best explained activity limitations after stroke. Physiotherapy Theory and Practice, 1–7.

11.

Michael,

K. M.

, Allen,

J. K.

, & Macko,

R. F.

(2005). Reduced ambulatory activity after stroke: the role of balance, gait, and cardiovascular fitness. Archives of Physical Medicine and Rehabilitation, 86(8), 1552–1556.

12.

Park,

K. H.

, Kim,

D. Y.

, & Kim,

T. H.

(2015). The effect of step climbing exercise on balance and step length in chronic stroke patients. Journal of Physical Therapy Science, 27(11), 3515–3518.

13.

Lee,

, & Seo,

(2014). The effects of stair walking training on the balance ability of chronic stroke patients. Journal of Physical Therapy Science, 26(4), 517–520.

14.

Fujita,

, Sato,

, Ohashi,

, Nishiyama,

, Ohashi,

, Yamane,

,... & Tozato,

(2018). Amount of balance necessary for the independence of transfer and stair-climbing in stroke inpatients. Disability and Rehabilitation, 40(10), 1142–1145.

15.

Bohannon,

R. W.

, & Walsh,

(1991). Association of paretic lower extremity muscle strength and standing balance with stair-climbing ability in patients with stroke. Journal of Stroke and Cerebrovascular Diseases, 1(3), 129–133.

16.

Polese,

J. C.

, Scianni,

A. A.

, & Teixeira-Salmela,

L. F.

(2015). Predictors of energy cost during stair ascent and descent in individuals with chronic stroke. Journal of Physical Therapy Science, 27(12), 3739–3743.

17.

Bonnyaud,

, Zory,

, Pradon,

, Vuillerme,

, & Roche,

(2013). Clinical and biomechanical factors which predict timed up and down stairs test performance in hemiparetic patients. Gait & posture, 38(3), 466–470.

18.

Polese,

J. C.

, Ribeiro-Samora,

G. A.

, Lana,

R. C.

, Rodrigues-De-Paula,

F. V.

, & Teixeira-Salmela,

L. F.

(2017). Energy expenditure and cost of walking and stair climbing in individuals with chronic stroke. Brazilian Journal of Physical Therapy, 21(3), 192–198.

19.

Modai,

, Sharon,

, Bar-Haim,

, & Hutzler,

(2015). Stair climbing test post-stroke: feasibility, convergent validity and metabolic, cardiac, and respiratory responses. Topics in Stroke Rehabilitation, 22(4), 281–288.

20.

Aziz,

A. R.

, & Teh,

K. C.

(2005). Physiological responses to single versus double stepping pattern of ascending the stairs. Journal of Physiological Anthropology and Applied Human Science, 24(4), 253–257.

21.

De Meersman,

R. E.

(1993). Heart rate variability and aerobic fitness. American Heart Journal, 125(3), 726–731.

22.

Shapira-Vadler,

, Treger,

, & Katz-Leurer,

(2015). Muscle Strength, Function and Heart Autonomic Regulation System Recovery at the Sub-Acute Stage Post Stroke. European Neurology, 74(3-4), 154–157.

23.

Raimundo,

R. D.

, de Abreu,

L. C.

, Adami,

, Vanderlei,

F. M.

, de Carvalho,

T. D.

, Moreno,

I. L.

,... & Sato,

M. A.

(2013). Heart rate variability in stroke patients submitted to an acute bout of aerobic exercise. Translational Stroke Research, 4(5), 488–499.

24.

Hsueh,

I. P.

, Lin,

J. H.

, Jeng,

J. S.

, & Hsieh,

C. L.

(2002). Comparison of the psychometric characteristics of the functional independence measure, 5 item Barthel index, and 10 item Barthel index in patients with stroke. Journal of Neurology, Neurosurgery & Psychiatry, 73(2), 188–190.

25.

Task Force of the European Society of Cardiology. (1996). Heart rate variability, standards of measurement, physiological interpretation, and clinical use. Circulation, 93, 1043–1065.

26.

Nunan,

, Jakovljevic,

D. G.

, Donovan,

, Hodges,

L. D.

, Sandercock,

G. R.

, & Brodie,

D. A.

(2008). Levels of agreement for RR intervals and short-term heart rate variability obtained from the Polar S810 and an alternative system. European Journal of Applied Physiology, 103(5), 529–537.

27.

Farah,

B. Q.

, Lima,

A. H. R. D. A.

, Cavalcante,

B. R.

, Oliveira,

L. M. F. T.

, Brito,

A. L. D. S.

, Barros,

M. V. G.

, & Ritti-Dias,

R. M.

(2016). Intra-individuals and inter-and intra-observer reliability of short-term heart rate variability in adolescents. Clinical Physiology and Functional Imaging, 36(1), 33–39.

28.

Cameron,

, & Bohannon,

R. W.

(2000). Criterion validity of lower extremity Motricity Index scores. Clinical Rehabilitation, 14(2), 208–211.

29.

Fayazi,

, Dehkordi,

S. N.

, Dadgoo,

, & Salehi,

(2012). Test-retest reliability of Motricity Index strength assessments for lower extremity in post stroke hemiparesis. Medical Journal of the Islamic Republic of Iran, 26(1), 27.

30.

Katz-Leurer,

, Fisher,

, Neeb,

, Schwartz,

, & Carmeli,

(2009). Reliability and validity of the modified functional reach test at the sub-acute stage post-stroke. Disability and Rehabilitation, 31(3), 243–248.

31.

Faria-Fortini,

, Basílio,

M. L.

, Polese,

J. C.

, Menezes,

K. K.

, Faria,

C. D.

, Scianni,

A. A.

, & Teixeira-Salmela,

L. F.

(2017). Strength deficits of the paretic lower extremity muscles were the impairment variables that best explained restrictions in participation after stroke. Disability and Rehabilitation, 39(21), 2158–2163.

32.

Raphaely,

B. N.

, & Katz-Leurer,

(2014). The Association between Cardiac Autonomic Control System and Motor Performance among Patients Post Stroke: Review of the Literature. Int J Neurorehabilitation, 1, 136.

33.

Bassi,

, Colivicchi,

, Santini,

, & Caltagirone,

(2007). Cardiac autonomic dysfunction and functional outcome after ischaemic stroke. European Journal of Neurology, 14(8), 917–922.

34.

Graff,

, Gasecki,

, Rojek,

, Boutouyrie,

, Nyka,

, Laurent,

, & Narkiewicz,

(2013). Heart rate variability and functional outcome in ischemic stroke: a multiparameter approach. Journal of Hypertension, 31(8), 1629–1636.

35.

Tang,

S. C.

, Jen,

H. I.

, Lin,

Y. H.

, Hung,

C. S.

, Jou,

W. J.

, Huang,

P. W.

,... & Jeng,

J. S.

(2015). Complexity of heart rate variability predicts outcome in intensive care unit admitted patients with acute stroke. J Neurol Neurosurg Psychiatry, 86(1), 95–100.

36.

Arad,

, Abboud,

, Radai,

M. M.

, & Adunsky,

(2002). Heart rate variability parameters correlate with functional independence measures in ischemic stroke patients. Journal of Electrocardiology, 35, 243.

37.

National acute stroke Israeli (NASIS), registry. (2013). National registry of hospitalized patients April-May 2013 preliminary findings.

38.

Shiber,

J. R.

, Fontane,

, & Adewale,

(2010). Stroke registry: hemorrhagic vs ischemic strokes. The American Journal of Emergency Medicine, 28(3), 331–333.

39.

Kaikkonen,

, Nummela,

, & Rusko,

(2007). Heart rate variability dynamics during early recovery after different endurance exercises. European Journal of Applied Physiology, 102(1), 79–86.

Characteristics		Sub-groups
	All N = 36	Independence N = 24	Dependence N = 12	p-value
Socio-demographic
Age (years)	56.36±12.01	59.5 [29.0–73.0]	62.0 [23.0–72.0]	0.57
Gender (men)	24 (66.7)	15 (62.5)	9 (75.0)	0.45
Health habits
Body mass index (kg/meter²)	27.36±5.22	25.7 [21.4–40.6]	25.8 [21.4–46.5]	0.97
Exercise (days a week)	1.22±1.79	1 [0–7]	0 [0–2]	0.02
Smoking (no)	33 (91.7)	2 (8.3)	1 (8.3)	1.00
Comorbidities
Hypertension	27 (75.0)	17 (70.8)	10 (83.3)	0.41
Hyperlipidemia	17 (47. 2)	12 (50.0)	5 (41.7)	0.64
Diabetes mellitus	11 (30.6)	5 (20.8)	6 (50.0)	0.07
Stroke
Hemorrhagic	20 (55.6)	15 (62.5)	5 (41.6)	0.48
Side weakness (left)	19 (54.3)	12 (52.2)	7 (58.3)	0.73