Repeated Treatments With Botulinum Toxin Type A Produce Sustained Decreases in the Limitations Associated With Focal Upper-Limb Poststroke Spasticity for Caregivers and Patients

Treatment 

As determined by the investigators, patients were treated with BTX-A (Botox) at a total dose of 200 to 400U, divided and injected intramuscularly. At least 200U was injected into the wrist and finger flexor muscles. Doses in excess of 200U were targeted to the flexed elbow, pronated forearm, thumb in palm, clenched fist, and intrinsic-plus hand deformity. Intrinsic-plus hand deformity results from overactivity of the interosseous, lumbrical, or hypothenar muscles that causes the fingers to stiffen and the hand to become deformed and functionally disabled. Treatment of adducted or internally rotated shoulders was not allowed. Retreatment was permitted at 12 or more weeks after the last treatment. To qualify for retreatment, the patient must have had, in the opinion of the investigator, sufficient spasticity in the wrist and finger flexors to warrant at least the minimum protocol treatment. The exact dosage and location of retreatments were at the discretion of the investigators and were not required to be identical to the initial treatment, but at least 200U had to be injected into the wrist and finger flexor muscles. Patients were stratified based on their BTX-A dosage into a high-dose (≥250U) and low-dose (200 to <250U) group.

Patients were allowed to receive physical and occupational therapies, other than casting, as prescribed by the treating physician. Any other prescription therapies (other than any additional botulinum toxin) or over-the-counter concomitant medications (including oral antispasticity medications) that were considered necessary could be given at the discretion of the investigators.

Assessment Instruments 

Assessments were performed before each treatment and every 6 weeks thereafter until week 54 or early exit from the study. Investigator- and patient-rated outcome measures were selected based on their validity and/or specificity for evaluating poststroke spasticity. Investigator-rated measures were the Disability Assessment Scale (DAS) and the Ashworth Scale to assess muscle tone. Patient-rated measures were the Stroke Adapted Sickness Impact Profile (SA-SIP30) and the visual analog scale (VAS) of the European Quality of Life−5 Dimensions (EQ-5D) health questionnaire. Staff at the study centers administered the questions of the SA-SIP30. Patients completed the EQ-5D.

Investigator-Rated Outcome Measures 

Patient-Rated Outcome Measures 

Safety Assessments 

Safety was assessed at each visit by the type and severity of the adverse events reported. Vital signs were also recorded at each visit. Physical examinations were performed at the screening visit (2 weeks before the first treatment) and at the exit visit. Blood was drawn at the screening and exit visits for complete blood cell counts, blood chemistry, and electrolyte determinations.

Serum Antibody Analysis 

Blood samples for detection of botulinum toxin–neutralizing antibodies were obtained before each treatment and at the exit visit. Samples were analyzed for neutralizing antibodies by using a mouse protection assay.29

Statistical Analysis 

Efficacy and safety analyses were performed in all patients who were given at least 1 treatment with BTX-A (intention-to-treat [ITT] population). All data were summarized by the total (overall) population, the low-dose group (BTX-A 200 to <250U), and the high-dose group (BTX-A ≥250U). Changes from baseline were evaluated by 95% confidence intervals or paired-samples t test. A P value of less than .05 was considered statistically significant. Unadjusted P values are reported. In addition, the closed-testing Hommel method30 was used in the analysis of the Ashworth Scale data to handle the multiplicity issue that arises from making multiple comparisons. Hommel adjustments were made on the basis of families of comparisons—for example, for the Ashworth Scale, separate Hommel adjustments were made for the wrist, finger, and shoulder. Given the limited number of thumb injections, Hommel adjustments were not performed for these data.

For each principal therapeutic intervention target, the mean DAS score changes from baseline were summarized with summary statistics. These scores were also summarized by patients in whom there was at least a 1-unit improvement. Changes from baseline in the Ashworth Scale were summarized by site of injection (ie, elbow, wrist, finger, thumb) and analyzed by BTX-A dose (ie, high and low). The VAS scores and SA-SIP30 transformed scores were summarized by using summary statistics for the baseline values and the changes from those values. Changes within groups were assessed by using paired-samples t tests.

All adverse events were categorized by incidence and severity in accordance with the standards defined in the Medical Dictionary for Regulatory Activities.31 Baseline values and the changes in those values in laboratory results and vital signs were summarized by summary statistics.

Demographics and Disposition 

Two hundred seventy-nine patients were enrolled, and 226 (81%) completed the year-long study (high-dose group, 175 [77%]; low-dose group, 51 [23%]) (fig 1). The mean follow-up was 358 days (range, 1–480d), and follow-up was greater than or equal to 54 weeks (378d) in 215 (77%) patients. Patients' baseline and demographic characteristics are shown in table 1. The demographic characteristics were similar between patients treated with high-dose and low-dose BTX-A.

View Large Image Download to PowerPoint

Fig 1. Subject disposition.

Table 1. Patient Characteristics

	Characteristic	ITT Population (N=279)
	Age (y)⁎	58±13
	≥60y, n (%)	128(46)
	Sex, n (%)
	Male	150(54)
	Female	129(46)
	Ethnicity, n (%)
	White	229(82)
	Black	35(13)
	Asian	4(1)
	Hispanic	8(3)
	Other	3(1)
	Height (cm)⁎	171±10
	Weight (kg)⁎	82±18
	Duration of spasticity (y)⁎	4.4±4.9

⁎ Mean ± SD.

The reasons for early discontinuation were adverse events in 15 (5%) patients (high-dose, 9; low-dose, 6), personal reasons in 14 (5%) patients (high-dose, 8; low-dose, 6), and other reasons in 12 (4%) patients (all high-dose). Among the 279 patients, most (84%) received 3 to 5 BTX-A treatments: 11% received 1 treatment, 6% received 2 treatments, 12% received 3 treatments, 22% received 4 treatments, and 49% received 5 treatments.

Investigator-Rated Outcomes 

Principal therapeutic intervention target and disability assessments 

One hundred thirty-eight (49%) of the 279 patients chose limb posture as the principal therapeutic intervention target, 82 (29%) chose dressing, 34 (12%) chose hygiene, and 25 (9%) chose pain. At each time point, the mean principal therapeutic intervention target scores were decreased (improved) from the baseline values in both BTX-A dose groups; the mean changes ranging from −1 to −0.6 in the high-dose group and from −0.9 to −0.5 in the low-dose group (table 2). The mean changes in the principal therapeutic intervention target scores in each dose group and in both dose groups combined were significantly lower at each time point. During the course of the study there was a decrease of at least 1 unit on the 4-point scale in 50% to 67% of the patients in the high-dose group and in 47% to 70% of those in the low-dose group and of at least 2 units in 18% to 31% of the patients in the high-dose group and in 8% to 28% of those in the low-dose group.

Table 2. DAS Principal Therapeutic Intervention Target: Mean Baseline and Change From Baseline, by BTX-A Dosage Group

	DAS, Principal Therapeutic Intervention Target	BTX-A 200 to <250U (n=66)	BTX-A ≥250U (n=213)
	Baseline	2.1±0.70	2.4±0.68
	Week 6	−0.5±0.78	−0.8±0.92
	Week 12	−0.6±0.96	−0.6±1.02
	Week 18	−0.8±1.02	−0.8±1.04
	Week 24	−0.8±0.97	−0.7±1.02
	Week 30	−0.9±0.94	−1.0±1.02
	Week 36	−0.9±0.96	−0.8±1.02
	Week 42	−0.9±1.05	−1.0±1.05
	Week 48	−0.9±0.97	−0.9±1.05
	Week 54	−0.8±0.96	−1.0±1.10

NOTE. Values are mean ± SD. DAS, Principal Therapeutic Intervention Target rated on a scale of 0 (no disability), 1 (mild disability), 2 (moderate disability), and 3 (severe disability).

Ashworth Scale 

Muscle tone, as assessed by the Ashworth Scale, was statistically significantly reduced in all muscle groups at all time points in the overall (ie, high and low) BTX-A dose groups in the ITT populations (table 3), even when adjusted for multiple comparisons. There was a suggestion of a dose effect in that the improvements in muscle tone in all muscle groups were greater in the high-dose group than in the low-dose group.

Table 3. Ashworth Scale Scores: Mean Baseline and Changes From Baseline in Treated Patients, by Muscle Groups

	Time	Elbow		Wrist		Finger		Thumb
		<250U	≥250U	<250U	≥250U	<250U	≥250U	<250U	≥250U
	Baseline	1.7±0.9	2.4±0.9	2.6±0.9	2.9±0.7	2.7±0.8	3.0±0.8	1.7±1.0	1.9±1.0
	6	−0.2±0.6⁎	−0.7±0.9	−0.8±1.0	−1.2±0.9	−0.7±1.0	−1.1±1.0	−0.3±0.8	−0.6±1.0
	12	−0.1±0.8†	−0.5±0.8	−0.5±1.1	−0.6±0.8	−0.5±0.8	−0.5±0.8	−0.2±0.7††	−0.3±1.0
	18	−0.3±0.9‡	−0.7±0.8	−1.0±1.0	−1.4±0.9	−1.0±1.0	−1.2±1.0	−0.5±0.9	−0.8±1.1
	24	−0.2±0.7§	−0.5±0.9	−0.7±1.0	−0.8±0.8	−0.6±0.9	−0.7±1.0	−0.2±1.0††	−0.4±1.2
	30	−0.3±0.9∥	−0.6±0.9	−1.2±1.1	−1.3±1.1	−1.0±1.0	−1.2±1.0	−0.5±1.1	−0.6±1.1
	36	−0.3±1.0¶	−0.6±1.0	−0.9±1.3	−1.1±1.1	−0.7±1.2	−0.9±0.9	−0.3±1.0	−0.6±1.1
	42	−0.1±0.9#	−0.7±0.9	−1.0±1.2	−1.3±1.1	−0.9±1.0	−1.2±0.9	−0.5±1.1	−0.7±1.2
	48	−0.3±0.8⁎⁎	−0.7±0.9	−1.1±1.1	−1.1±1.1	−0.8±1.0	−0.8±1.0	−0.5±1.1	−0.6±1.2
	54	−0.3±0.7	−0.8±1.0	−1.1±1.2	−1.4±1.0	−0.9±1.0	−1.3±1.0	−0.3±1.0	−0.8±1.1

NOTE. Values are mean ± SD. The flexor tone was assessed using the 5-point Ashworth Scale (0 [none, no increase in muscle tone] to 4 [very severe, limb rigid in flexion or extension]). All changes from baseline are significant for Hommel adjusted (P<.01). P values, except as noted for Hommel P values, are below.

⁎ P=.147. † P=.511. ‡ P=.072. § P=.086. ∥ P=.153. ¶ P=.156. # P=.511. ⁎⁎ P=.086. †† Not determined.

As summarized in table 3, injection of BTX-A in the wrist, finger, and thumb was associated with significant improvements from baseline in muscle tone at each study time point in the low- and the high-dose groups and overall, with the exceptions of low-dose thumb injections at weeks 12 (unadjusted P=.055) and 24 (unadjusted P=.139). Significant improvements from baseline in muscle tone were also noted in those receiving elbow treatment; however, these improvements were consistent in only the high-dose group and overall population. A significant improvement from baseline was noted in the low-dose elbow group at week 54. Those treated in the elbow using low-dose BTX-A showed statistically significant improvement in the nonadjusted analysis. However, when corrected for multiple comparisons, these differences were nonsignificant at weeks 6, 18, 24, 48, and 54. There was a trend toward significance at weeks 30 and 36 (P=.051, P=.052) in the nonadjusted analysis. These differences, of course, washed out when multiplicity was taken into consideration.

Patient-Rated Outcomes 

SA-SIP30 

The mean changes in the baseline SA-SIP30 scores in the physical and psychosocial dimensions, as well as the overall score, are shown in table 4. The mean ± standard deviation (SD) physical dimension baseline score in the total study population was 49.2±24.3. The mean scores declined (improved) at week 6 and week 12 in both treatment groups, but the changes were not significant (median changes were generally 0). At week 24 there was a statistically significant reduction in the mean score ± SD in the entire group (−2.9±16.1, P=.007). The mean scores showed a trend for improvement (decline in score), and the differences from the baseline score were statistically significant at week 30 (−2.5, P=.024), week 36 (−3.3, P=.003), and week 48 (−2.6, P=.048). The mean decreases ± SD in the high-dose group were statistically significant at weeks 24 (−3.0±16.6, P=.017) and 36 (−3.6±16.0, P=.004), but those in the low-dose group were not statistically significant at any time point.

Table 4. Summary of SA-SIP30 Baseline and Changes From Baseline in Treated Patients at Each Follow-Up Visit

	Visit	Overall Score	Physical Dimension	Psychosocial Dimension
	Baseline	41.5±19.9	49.2±24.3	32.4±22.9
	Change at week
	6	−2.1±11.9⁎	−1.1±14.2	−3.4±16.3⁎
	12	−2.5±12.6⁎	−1.4±16.5	−3.9±14.7⁎
	24	−2.9±13.3⁎	−2.9±16.1⁎	−2.9±17.1⁎
	30	−3.5±13.3⁎	−2.5±16.7⁎	−4.7±17.6⁎
	36	−4.0±13.5⁎	−3.3±16.4⁎	−4.8±17.2⁎
	48	−3.5±14.3⁎	−2.6±17.7⁎	−4.6±18.3⁎

NOTE. Values are mean ± SD. The SA-SIP30 was used to measure 2 dimension scores (physical, psychosocial) and a global score on a scale of 0 (no dysfunction) to 100 (maximum dysfunction).

⁎ P≤.05 for change from baseline.

The mean baseline psychosocial dimension score ± SD in the entire study population was 32.4±22.9. There were statistically significant reductions in the mean psychosocial scores at all time points in the study (P≤.01). At week 6 the mean scores improved by −3.4 (P<.001), and the improvements ranged from −2.9 to −4.8 at the follow-up visits. At the conclusion of the study the mean scores had significantly improved, by −4.6 points (P<.001). There were improvements in both groups, but the improvements were statistically significant at all time points only in the high-dose group.

The mean overall baseline score ± SD was 41.5±20.0. There were statistically significant improvements in the mean overall score at all time points in the study. The mean overall score improved by −2.1 (P=.004) at week 6, with continued improvements over time, ranging from −2.1 to −4.0. At the conclusion of the study the mean overall score had improved −3.5 points (P=.001). There were improvements in both dose groups, but improvements were statistically significant at all time points only in the high-dose group and in the entire population.

VAS of the EQ-5D 

Patients rated their health statuses by using the VAS of the EQ-5D. The mean baseline health state score was 66.4±21.3 in the overall group, 66.7±21.8 in the high-dose group, and 65.7±20.1 in the low-dose group (0, worst imaginable health; 100, best imaginable health). Increases in the health state score from baseline were observed in both groups (table 5) with mean changes from baseline ranging from 0.5 to 4.2 in the overall group, 0.8 to 4.4 in the high-dose group, and −0.5 to 4.1 in the low-dose group.

Table 5. Baseline and Change From Baseline in HRQOL Questionnaire

	Health State Score	BTX-A 200 to <250 U (n=66)	BTX-A ≥250 U (n=213)
	Baseline	65.7±20.1	66.7±21.8
	Week 6	66.2±20.4⁎	67.8±20.3⁎
	Week 12	67.2±18.2⁎	69.9±19.7⁎
	Week 24	67.3±17.3⁎	70.6±20.2⁎
	Week 30	70.9±18.1⁎	72.7±19.2⁎
	Week 36	71.0±17.9⁎	72.3±19.6⁎
	Week 48	70.1±16.6⁎	73.1±19.2⁎

NOTE. Values are mean ± SD. Health State Score measured on a scale from 0 (worst imaginable health state) to 100 (best imaginable health state).

⁎ P<.001 compared with baseline (paired t test).

Safety 

Adverse events were reported in 77% (216/279) of all patients—79% (169/213) in the high-dose group and 71% (47/66) in the low-dose group, with the vast majority not considered related to treatment as discussed below. The rates of the specific adverse events were similar in the 2 dose groups. Adverse events that were reported in at least 5% of patients were upper-respiratory tract infection (13% [37/279]), urinary tract infection (11% [30/279]), pain in extremity (9% [24/279]), fall (8% [23/279]), arthralgia (7% [19/279]), headache (6% [17/279]), and peripheral edema (6% [16/279]). Serious adverse events were reported in 49 patients (18%; n=39, ≥250U; n=10, 200 to <250U). All serious adverse events were considered by the investigators to be unrelated to the study medication.

Treatment-related adverse events were reported in 6% (18/279) of all patients—7% (14/213) in the high-dose group and 6% (4/66) in the low-dose group. Treatment-related adverse events that were reported in at least 3 patients were muscle weakness (high-dose group, 4; low-dose group, 2) and pain in the extremity (high-dose group, 4), all of which were mild or moderate. The only treatment-related adverse event that resulted in withdrawal from the study was muscle weakness in the fingers, which occurred in 1 patient in the high-dose group.

There were no clinically relevant changes from baseline values in any chemistry or hematology test results, pulse rate, systolic or diastolic blood pressure, or findings on physical examination.

Toxin-Neutralizing Antibodies 

Posttreatment serum samples sufficient for analysis for toxin-neutralizing antibodies were available in 80% (224/279) of patients, with negative test results in 221 patients, inconclusive results in 2, and positive results in 1 (<0.5%). The patient with a positive result had no history of botulinum toxin treatment and had a negative test result at baseline. During the study, the patient received 5 treatments of 280 to 400U, each 12 weeks apart. After the first 3 treatments, laboratory test results were negative for neutralizing antibodies, and the patient responded to treatment. Responses were diminished with the final treatments; the patient tested positive for neutralizing antibodies before the fifth treatment. Resistance to botulinum toxin was confirmed with the frontalis type A test.32

Study Limitations 

Study limitations include issues related to the enrollment of patients with prior BTX-A exposure. Patients were excluded from this trial if they had received BTX-A in a clinical study for spasticity at any time, for any reason before 1998, or within 4 months of the study or if they had prior exposure to another formulation of botulinum toxin. Patients were allowed to participate in this study provided they had not received treatment at least 4 months before study enrollment. Although it is generally accepted that BTX-A lasts 3 to 6 months in patients treated with poststroke spasticity, there may have been some patients enrolled who had not returned to baseline at the time of enrollment. Another potential limitation is the use of a subjective measure of functional improvement, the DAS. Although supported by a previous investigation that showed its interrater and intrarater reliability,17 by virtue of its subjective nature there are limitations to the conclusions that may be drawn using this instrument. In addition, the measurement of muscle tone using the Ashworth Scale, although also found to have good intrarater and interrater reliability,17 may also be limited by its subjective nature. In addition, because tone can vary with many things, such as time of day and patient's emotional state, it is possible that the Ashworth changes reported may also reflect other changes outside of the intervention with BTX-A.

Treatment with BTX-A was well tolerated in this population. Up to 5 treatments were administered, and there were few treatment-related adverse events, all of which were mild or moderate and were transient. Only 1 patient (0.5% of study population) in this trial developed toxin-neutralizing antibodies after repeated treatment (ie, before the fifth treatment). Repeated exposure to BTX-A, especially in doses higher than 300U, has previously been shown to increase the risk of antibody development.39, 40 However, the development of toxin-neutralizing antibodies with BTX-A is rare. In a recent analysis of 895 samples from 6 studies in which patients were treated between 1 and 14 times with BTX-A for headache, cervical dystonia, or spasticity (the current study), only 4 samples tested positive for toxin-neutralizing antibodies.41 The relationship between the development of these antibodies and clinical response has been poorly correlated historically42; however, in the current study the presence of neutralizing antibodies was associated with decreased efficacy in the 1 patient in whom they were found. Overall, the low incidence of antibody development in the study population suggests that the efficacy of repeated therapy is not typically diminished over a 54-week period.

Considering that treatment with BTX-A affects only 1 aspect of the potential array of physical problems that patients face after stroke (not including the existing cognitive, visuospatial, sensory, and other central motoneuron deficits), the degree of improvement we found in functional disability, HRQOL, and muscle tone of patients over 1 year is notable. The results of this open-label, year-long study merit further exploration in a larger randomized controlled trial investigating comprehensive measures of repeated BTX-A treatment in poststroke spasticity patients.

The study was designed as an open-label study because the main goal was to assess the safety and clinical effects of repeated treatment with BTX-A over a prolonged period. This open-label design, however, limits the study because of possible subject and investigator bias, especially with subjects who are not toxin naive. In addition, in the absence of a placebo control group, interpretations of the magnitude of effect are limited.