Critical evaluation of the modified Rankin Scale for assessment of the efficacy of mechanical thrombectomy: A retrospective comparison between the modified Rankin Scale and functional independence measure

Abstract

Objective

The primary outcome measure used in mechanical thrombectomy (MT) trials is the modified Rankin Scale (mRS). However, the accuracy of mRS might be limited. On the other hand, the functional independence measure (FIM) is a widely used tool to quantify the extent to which patients require assistance during their activities of daily living. The current study aimed to reveal different clinical backgrounds that affect the efficacy of MT measured either by mRS or FIM.

Methods

Patients who underwent MT at our institution from January 2019 to July 2022 were included and divided into groups based on mRS scores of 0–2 and ≥ 3. Patients were also divided into two groups based on a cut-off value of FIM of ≥ 108, as patients with FIM ≥ 108 are capable of living an independent life.

Results

The mRS score was 0–2 in 33% of the patients, while the FIM score was ≥ 108 in only 15% of the patients. In the mRS groups, there were significant differences in terms of duration of hospitalization, National Institutes of Health Stroke Scale (NIHSS) scores, achievement of thrombolysis in cerebral infarction (TICI) reperfusion grade of 2b or 3, and postoperative bleeding. Multivariate logistic regression analysis revealed that NIHSS score and achievement of TICI 2b or 3 were significant factors related to mRS 0–2 at discharge. The FIM groups differed significantly in terms of age and, duration of hospitalization, NIHSS score, although multivariate logistic regression analysis revealed that only the NIHSS score was significantly associated with an FIM score of ≥ 108.

Conclusion

The study showed that the percentage of independent patients is significantly reduced when we evaluated the patients by the FIM. In addition, there are some differences in the clinical background that led to a good outcome between that evaluated by mRS and FIM.

Introduction

Many randomized clinical trials have proved the efficacy of mechanical thrombectomy (MT) for acute ischemic stroke (AIS) due to large vessel occlusion.^1,2 The primary outcome measures used for these trials were the modified Rankin Scale (mRS) at 90 days, and cerebral hemorrhagic events, and mRS 0–2 was defined as a good outcome.^3,4 Although the mRS is the “de facto standard” outcome measure for stroke clinical trials, its reproducibility has been questioned and discussed. ⁵ Additionally, the accuracy of mRS for evaluating the patient's condition might be limited, as the score only has a range of seven levels, from 0 to 6.

The functional independence measure (FIM) is another widely used tool to quantify the extent to which patients require assistance in their activities of daily living. ⁶ The FIM instrument consists of 18 items that measure each individual's functional disability in terms of their need for assistance. The FIM instrument measures the two unidimensional domains of motor function (13 items) and cognitive function (fives items),^7,8 with each item scored from 1 to 7 (Table 1).

Table 1.

Functional independence measure (FIM).

Component of functional independence measure
Motor component	Cognitive component
Self-care	Comprehension
Eating	Expression
Grooming	Social interaction
Bathing	Promlem solving
Dressing upper body	Memory
Dressing lower body
Toileting
Sphincter control
Bladder management
Bowel management
Transfers
Bed/chair/wheelchair
Toilet
Bath/shower
Locomotion
Walking/wheelchair
Stairs
Score of functional independence measure
Degree of dependency	Level of functioning
No helper	7  =  Complete independence
	6  =  Modified independence
Modified dependence on a helper	5  =  Supervision
	4  =  Minimal assistance
	(at least 75% independent)
	3  =  Moderate assistance
	(at least 50% independent)
Complete dependence on a helper	2  =  Maximal assistance
	(at least 25% independent)
	1  =  Total assistance
	(< 25% independent)

The current study first aimed to compare the observed effectiveness of MT for AIS when using the FIM versus mRS and then investigated if any difference in the clinical background led to a good outcome between that evaluated by mRS and FIM.

Methods

Cohort design

Our local institutional review board approved the use of the patient's clinical data for this research (approval number: 2128). Patients who received MT at our institution from January 2019 to July 2022 were included. AIS cases due to lesions in the posterior circulation (four cases), cases lacking sufficient data (four cases), and cases whose FIM evaluation date differed by seven days or more from the date of discharge (16 cases) were excluded from analysis (Figure 1).

Figure 1.

Inclusion of patients in the study cohort.

Clinical variables

The clinical characteristics of the entire cohort are shown in Table 2. The clinical variables collected from the patient's medical records included hypertension, diabetes mellitus, history of stroke, prehospital mRS score, administration of recombinant tissue-type plasminogen activator (rt-PA), cardiogenic infarction, and duration of hospitalization. The presence of hypertension was defined as a history of antihypertensive therapy or of being diagnosed with hypertension by a doctor. The presence of diabetes mellitus was defined by an HbA1c level greater than 6.4% or a history of anti-diabetic drug therapy. National Institutes of Health Stroke Scale (NIHSS) scores were assessed at admission, and Diffusion-Weighted Imaging–Alberta Stroke Program Early Computed Tomography Score (DWI-ASPECTS) was used for evaluation of preoperative magnetic resonance images (MRI). ⁹ Postoperative hemorrhage was evaluated by MRI and computed tomography (CT) according to the criteria of the European-Australasian Acute Stroke Study (ECASS II), where parenchymal hemorrhage 1 (PH1) and PH2 were considered to indicate the presence of postoperative hemorrhage. ¹⁰

Table 2.

Characteristics of the patients (n  =  86).

Mean age (years) (mean  ±  SD)	77.7  ±  10.0
Gender (women: men)	43:43
Hypertension	47 (55%)
Diabetes mellitus	18 (19%)
History of stroke	15 (17%)
Prehospital mRS 0–2	82 (95%)
Affected side (right: left)	37:49
Cardiogenic	71 (83%)
Hospitalization days	27.5 (18.0–36.8)
NIHSS (mean  ±  SD)	16.8  ±  6.9
rt-PA	28 (33%)
DWI-ASPECTS	7.0  ±  2.1
TICI 2b or 3	67 (78%)
Postoperative hemorrhage	15 (17%)
Diseased vessel
ICA	22 (26%)
M1	55 (64%)
M2	9 (10%)

SD: standard deviation; mRS: modified Rankin Scale; NIHSS: National Institutes of Health Stroke Scale; rt-PA: recombinant tissue-type plasminogen activator; TICI: thrombolysis in cerebral infarction; ICA: internal carotid artery.

MRs and FIM scoring to evaluate the patient's independence in daily life

The mRS is the most used functional outcome measure in contemporary stroke scales, in which the outcome is classified into seven levels.^11,12 We divided patients into groups based on mRS scores of 0–2 and ≥ 3, as most clinical trials use mRS scores of 0–2 as a favorable outcome, assuming that patients are in a near-independent state.^1,3,13 On the other hand, the FIM is a widely used tool to quantify the extent to which patients require assistance with their activities of daily living. ⁶ The FIM instrument scores disabilities for 18 different items from 1 to 7, with a score of 6 or 7 indicating independence for the evaluated item by definition (Table 1). ¹⁴ As a result, a total score of 108, calculated by multiplying six by 18, suggests that the patient is entirely independent in their daily life. Thus, patients were divided into two groups with an FIM score of 108 as the cut-off value. All FIM scores were measured by physical therapists using the Japanese version of FIM™ version 3.0., ¹⁵ which has some relevant cultural modifications as compared to the original version.^14,16 FIM score was independently assessed by an occupational therapist (O.T.). Its evaluation is independent.

Statistical analysis

Comparative two-group analysis was performed using GraphPad Prism 9 (GraphPad Software, San Diego, California, USA) statistical analysis software. Statistical analysis was performed using Fisher's exact test to assess the associations between categorical variables. Distributed continuous variables were compared using the Student’s t-test and Mann-Whitney U-test. Multivariate logistic regression analysis was performed to identify significant factors by selecting variables from univariate analysis. A p-value of < 0.05 was considered statistically significant.

Results

Patients’ characteristics

Table 2 summarizes the characteristics of the 86 analyzed patients. There were 43 women and 43 men, with a mean age of 77.7  ±  10.0 years (± standard deviation). The number of patients with a history of hypertension, diabetes mellitus, and history of stroke was 47 (55%), 18 (19%), and 15 (17%), respectively. The number of patients with a prehospital mRS score of 0–2 was 82 (95%). The lesion was on the right side in 37 patients and on the left side in 49 patients. There were 71 cardiogenic ischemic stroke patients (83%). The mean NIHSS score was 16.8  ±  6.9 (± standard deviation). The number of patients given rt-PA was 28 (33%). The mean value of DWI-ASPECTS was 7.0  ±  2.1 (± standard deviation). Sixty-seven patients (87%) achieved thrombolysis in cerebral infarction (TICI) 2b or 3 after MT. The mean duration of hospitalization was 27.5 (18.0–36.8) days. The mRS and FIM scores don’t show identical distributions in the patients in this study (Figure 2). In particular, the variability of mRS 2 and 3 is remarkable. Furthermore, we noticed that patients with high cognitive-FIM scores tended to be scored favorably in mRS, even in cases with low motor-FIM scores (Figure 3). The mRS score was 0–2 in 33% of the patients, while the FIM score was ≥ 108 in only 15% of the patients (Figure 4).

Figure 2.

FIM score was plotted as a function of mRS. FIM and mRS scores showed similar data distributions but were not identical.

mRS: modified Rankin Scale; FIM: functional independence measure.

Figure 3.

Cognitive-functional independence measure (FIM) scores are plotted as a function of motor-FIM for patients with mRS 0–2. Datapoints are categorized by total FIM ≥ 108 (black circles) or <109 (crosses) with red crosses having high cognitive-FIM (>25) but low motor-FIM (<75).

Figure 4.

The frequency of “desirable” outcomes was compared between mRS and FIM scores. Thirty-three percent of the patients had mRS scores of 0–2, while only 15% of the patients had FIM scores of ≥ 108.

mRS: modified Rankin Scale; FIM: functional independence measure.

Factors corresponding to a desirable outcome (MRs 0–2) at discharge

Table 3 compares mRS scores 0–2 with a score of ≥ 3. There was no difference between the two groups in terms of age (p  =  0.78), gender (p > 0.99), presence of hypertension (p  =  0.82), presence of diabetes mellitus (p  =  0.40), number of cases with mRS score of 0–2 before onset (p > 0.99), affected side (p  =  0.10), cardiogenic stroke as the etiology (p  =  0.37), rt-PA administration (p  =  0.22), and DWI-ASPECTS (p  =  0.42). However, univariate analysis showed a significant difference between the two groups in terms of NIHSS score (p < 0.01), achievement of TICI 2b or 3 (p  =  0.03), postoperative bleeding (p  =  0.03), and duration of hospitalization (p < 0.01) on univariate analysis. Subsequent multivariate logistic regression analysis revealed that NIHSS score (p < 0.01; adjusted odds ratio: 1.16, 95% confidence interval (CI): 1.07–1.28) and achievement of TICI 2b or 3 (p  =  0.02; adjusted odds ratio: 0.14, 95% CI: 0.02–0.61) showed a significant correlation with an mRS score of 0–2 at discharge (Table 4), although postoperative hemorrhage (p  =  0.18) did not significantly correlate with an mRS score of 0–2 at discharge. The variance inflation factor (VIF) showed a low tendency for multicollinearity (age: 1.08, TICI 2b or 3: 1.00, postoperative hemorrhage: 1.09).

Table 3.

Comparison of cases with mRS scores of 0-2 and ≥3 at discharge.

	mRS 0-2 (n  =  28)	≥ 3 (n  =  58)	p-value
Mean age (years) (mean  ±  SD)	77.3  ±  9.9	77.9  ±  10.3	0.78
Gender (woman)	14 (50%)	29 (50%)	> 0.99
Hypertension	16 (57%)	31 (53%)	0.82
Diabetes mellitus	4 (14%)	14 (24%)	0.40
History of stroke	3 (10%)	12 (21%)	0.37
Prehospital mRS 0-2	27 (96%)	55 (95%)	> 0.99
Affected side (left)	12 (43%)	37 (64%)	0.10
Cardiogenic	25 (89%)	46 (79%)	0.37
NIHSS (mean  ±  SD)	12.5  ±  6.1	18.8  ±  6.4	< 0.01*
rt-PA	12 (43%)	16 (28%)	0.22
DWI-ASPECTS	7.3  ±  2.0	6.9  ±  2.1	0.42
TICI 2b or 3	26 (93%)	41 (71%)	0.03*
Postoperative hemorrhage	1 (0%)	14 (24%)	0.03*
Days of hospitalization	18.0 (10.0–23.5)	34.0 (24.5–39.5)	< 0.01*

SD: standard deviation; mRS: modified Rankin Scale; NIHSS: National Institutes of Health Stroke Scale; rt-PA: recombinant tissue-type plasminogen activator; TICI: thrombolysis in cerebral infarction.

*Variables showing significant differences.

Table 4.

Multivariate analysis of factors corresponding to mRS scores of 0-2 at discharge.

	OR	95% CI	p-value	VIF
NIHSS	1.16	1.07–1.28	< 0.01*	1.08
TICI 2b or 3	0.14	0.02–0.61	0.02*	1.00
Postoperative hemorrhage	4.40	0.69–86.2	0.18	1.09

NIHSS: National Institutes of Health Stroke Scale; TICI: thrombolysis in cerebral infarction; OR: odds ratio; CI: confidence interval; VIF: variance inflation factor.

*Variables showing significant differences.

Factors corresponding with an FIM score of ≥ 108 at discharge

Table 5 compares cases with FIM scores of ≥ 108 and < 108 at discharge. There was no difference between the two groups in terms of gender (p > 0.99), presence of hypertension (p > 0.99), presence of diabetes mellitus (p  =  0.29), number of cases with mRS score of 0–2 before onset (p > 0.99), affected side (p  =  0.54), cardiogenic stroke as the etiology (p  =  0.69), rt-PA administration (p  =  0.34), and DWI-ASPECTS (p  =  0.06). However, there was a significant difference between the two FIM groups in terms of age (p  =  0.03), NIHSS score (p < 0.01), and duration of hospitalization (p  =  0.03) on univariate analysis. Subsequent multivariate logistic regression analysis revealed that although the NIHSS score (p < 0.01; adjusted odds ratio: 0.75, 95% CI: 0.62–0.86) was significantly associated with an FIM score of ≥ 108 at discharge, age (p  =  0.37; adjusted odds ratio: 0.97, 95% CI: 0.91–1.04)) was not a significant factor for FIM score ≥ 108 at discharge, and VIF showed a low tendency for multicollinearity (age: 1.06, NIHSS: 1.06) (Table 6).

Table 5.

Comparison of cases with FIM scores of ≥ 108 with those with scores of < 108 at discharge.

	FIM≧108 (n  =  13)	FIM < 108 (n  =  73)	p-value
Mean age (years) (mean  ±  SD)	72.2  ±  12.3	78.7  ±  9.4	0.03*
Gender (woman)	6 (46%)	37 (51%)	>0.99
Hypertension	7 (54%)	40 (55%)	>0.99
Diabetes mellitus	1 (8%)	17 (23%)	0.29
History of stroke	2 (15%)	13 (18%)	>0.99
Prehospital mRS 0-2	13 (100%)	69 (95%)	>0.99
Affected side (left)	6 (46%)	43 (59%)	0.54
Cardiogenic	10 (77%)	61 (84%)	0.69
NIHSS (mean  ±  SD)	8.2  ±  4.1	18.3  ±  6.2	< 0.01*
rt-PA	6 (46%)	22 (30%)	0.34
DWI-ASPECTS	8.0  ±  1.6	6.8  ±  2.1	0.06
TICI 2b or 3	12 (92%)	55 (75%)	0.28
Postoperative hemorrhage	0 (0%)	15 (21%)	0.11
Days of hospitalization	18.0 (10.0–21.5)	31.0  ±  (21.0–31.0)	< 0.01*

SD: standard deviation; mRS: modified Rankin Scale; NIHSS: National Institutes of Health Stroke Scale; rt-PA: recombinant tissue-type plasminogen activator; TICI: thrombolysis in cerebral infarction; FIM: functional independence measure.

*Variables showing significant differences.

Table 6.

Multivariate analysis of factors corresponding to an FIM score of ≥ 108 at discharge.

	OR	95% CI	p-value	VIF
Age	0.97	0.91–1.04	0.37	1.06
NIHSS	0.75	0.62–0.86	< 0.01*	1.06

NIHSS: National Institutes of Health Stroke Scale; TICI: thrombolysis in cerebral infarction; OR: odds ratio; CI: confidence interval; VIF: variance inflation factor; FIM: functional independence measure.

*Variables showing significant differences.

Discussion

MT is now the standard treatment for ischemic stroke, with its treatment indications expanding year by year; it is believed to be beneficial for many stroke patients.^3,17 While the mRS score is a commonly used outcome evaluated in clinical trials on MT for AIS, it is not sensitive enough to detect small changes in the patient's condition. ¹⁸ Accurate assessment of the patient's daily activities might be difficult if small changes cannot be captured. Although FIM is a more complicated tool for evaluating outcome measures, it is thought to more thoroughly evaluate the patient's activities of daily life than the mRS, as FIM covers cognitive function assessment, which mRS does not.

For the current investigation, an FIM score of 108 or above was assumed to be equivalent to an mRS score of 0 to 2, referring to a previous report comparing FIM and mRS for motor function evaluation. ¹⁸ In the FIM, 13 items are included in the motor domain, which has a maximum possible score of 91 (7  ×  13), and the previous report showed that 95% of mRS 0, 100% of mRS 1, and 75% of mRS 2 patients had an FIM score of 80 or higher. The cut-off value will increase from 79 to 108 when this concept is expanded from 13 to 18 items.

As can be appreciated in Figure 2, the FIM score exhibits a higher resolution for evaluating patients’ conditions than the mRS score. Furthermore, mRS tended to be scored high tended to be scored higher if the patient's cognitive function was well preserved, even if their motor function substantially deteriorated (Figure 3), presumably due to the neurosurgeon's overestimation of the patient's function obscured by the well-preserved patient's communication skill. As a result, the percentage of patients considered “independent” reduced to half when they were evaluated by FIM (Figure 4). This result aligns with the “real-world” experience for the caregivers and patients where it is difficult to interpret the treatment as “meaningful” when the mRS evaluates it as successful. However, caution should be taken as FIM cannot be readily estimated before hospitalization. FIM requires intensive cognitive assessments, which could vary among patients, especially for elderly patients before disease onset while estimating prehospitalization-mRS is straightforward.

We were also able to elucidate those key clinical backgrounds contributing to good outcomes differed depending on whether mRS or FIM was used as the outcome measure and that NIHSS score at admission and achievement of TICI 2b or 3 were essential for discharging patients with an mRS score of 0–2. However, we found that age and NIHSS score at admission were significant factors related to an FIM score of ≥ 108 by univariate analysis. Multiple regression analysis suggested that the NIHSS score was the only significant factor related to an FIM score of ≥ 108 at discharge. Although age was not significantly associated with FIM ≥ 108 by multiple regression analysis, it is noteworthy that univariate analysis showed a significant difference in age between FIM ≥ 108 and < 108 groups (Table 5), since more studies now claim the efficacy of MT for elderly patients when using the mRS as the primary outcome measure.^19,20 Our results suggest that the significance of age as a contributing factor to MT efficacy might be underestimated if the treatment outcome is measured using only the mRS, while the effect of age cannot be ignored when the efficacy of MT is assessed using the FIM. Thus, the indication for offering MT to elderly AIS patients should be determined with great care, considering the age of the patient, NIHSS score at admission, and ischemic core volume, similar to what was followed in the DAWN trial. ¹⁷

Some limitations regarding the current study should be addressed. First, mRS and FIM scores were only evaluated at discharge, although ideally, they should be evaluated at different time points after the procedure, such as at 30-, 90-, and 180-days post-discharge. In fact, the duration of hospitalization differed between groups with different mRS or FIM scores. The results might differ when FIM and mRS scores are prospectively assessed at prespecified time points. Second, using FIM ≥ 108 as the cut-off for favorable outcomes is debatable, and further external validation is warranted for using this cut-off value as an outcome measure. Third, only ischemic occlusions affecting the internal carotid artery and M1 and M2 middle cerebral artery were evaluated. The results could be different for posterior circulation occlusions. Lastly, there were more left-side affected cases (49) than the right side (37, Table 2). Although the affected side did not significantly influence FIM (Table 5), we cannot ignore that more left-side cases had some impact on the obtained result.

Conclusions

When comparing mRS and FIM scores to evaluate the efficacy of MT, the definition of independence in daily life activities might be stricter with the FIM score, resulting in different results on the efficacy of MT. Moreover, patient age might affect the results of the assessment of the FIM score.