3. Discussion
We demonstrated in this study that the outcome of BoNT treatment for post-stroke CFD depends on the effect of FDL on the individual toes; the treatment was more effective when contraction of the FDL muscle controlled the movement of more than one toe, such as the second (the tibial side) to fifth toe. BoNT treatment of the CFD was more effective 2.5 to 6.0 times when the action of the FDL muscle extended by one toe on the tibial side. This finding was statistically adjusted for the confounding factor of causative disease.
The FDL muscle is small compared to the space/volume of the lower leg, and for this reason an electrical stimulator, an electromyogram (EMG) measuring device, or an echo device is required to reliably administer the drug to the muscle belly. In recent years, it has become more common to administer drugs in a more non-invasive manner using echo devices. In this case as well, if there is movement in the cross-section of the muscle on the screen when each toe is passively moved one by one in order while observing the FHL and FDL muscles with an echo device, it can be inferred that the muscle is transmitting tension to that toe. Therefore, this is a clinically useful method of administering drugs without using an electrical stimulator. We have performed many botulinum toxin treatments for CFD, and we have experienced only one side effect of excessive extension of the first toe after administering the drug to the FHL muscle, but we have never experienced such a side effect with the FDL muscle.
In the present study, we used two indices to classify the mode of control of the FHL and FDL muscles on each toe. MCT (the mode of control over the toes) is a qualitative variable that provides ordinal scale in the individual patient. In other words, it is a single index that reflects the manner in which the FHL or FDL muscle controls the toe movement. On the other hand, the NSM (the number of strongly moving toes) is a quantitative variable that represents the number of patient’s toes that moved in response to each BoNT injection. The same patient may show variable response each time in multiple drug administrations of BoNT, and thus the same patient can have several different numbers of these indices. It seems that MCT rather than NSM is related to the effectiveness of the treatment. Thus, rather than focusing on inducing more toe movements with electrical stimulation of the FDL muscle during BoNT injection, the patient-specific mode of dominance of the FDL muscle over the individual toes may determine the outcome of BoNT treatment in post-stroke patients with CFD. In this study, we could not find any specific factor that can be controlled by the therapist and that influenced the outcome of treatment. We could not elucidate how to treat patients whose FDL muscles control only the fewer toes to obtain a better therapeutic effect. Future studies need to be more objective and detailed in their assessment of the effects of BoNT treatment in this population of patients. Furthermore, regarding long-term treatment strategies, we may suggest the administration dose over time of the drug.
Although they had no effect on the conclusions of this study, analyzing the relationship of NSM of the FHL muscle and the causative disease and treatment outcome suggested its multicollinearity, probably due to the association between the underlying condition and the treatment outcome in patients with documented NSM of the FHL. We presume the multicollinearity is related to the effects of the less-than-ideal methodology used to determine the treatment outcome and the relatively small number of patients.
Toe movement varies each time with multiple electrical stimulations of the FHL and FDL muscles. In order to investigate such variability in the present study, we used the term “variability”, which reflected the standard deviation of the number of toes that moved following muscle contraction within the same subject (= SD-NSM). Therefore, the standard for “variability” is 0 if movement occurred in the same toe following each muscle contraction, and 0.7 if movement included adjacent toes. During electrical stimulation of the FHL muscle, the above variability index (SD-NSM) in the same subject was 0.53, while that following electrical stimulation of the FDL muscle SD-NSM was 0.52. These results suggest that the range of “variation” for both the FHL and FDL muscles is generally the adjacent toe or less. The within-subject variability was modest.
Since the FHL muscle branches out to the FDL muscle as well as to the big toe, we believe that treatment of the second and third CFDs with BoNT should involve the injection of BoNT into the FHL muscle as well as the FDL muscle [
1,
2]. In addition to it, we now recognize that the more toes that move with FDL muscle stimulation, the better the outcome would be with the BoNT treatment, giving us a better prediction of the treatment outcome, as well as enabling us to avoid unnecessary drug injections to the patients.
In this study, we categorized the dominant type of the FHL and FDL muscles for each toe by analyzing the results of multiple sessions in patients who underwent repeated electrical stimulations. The results showed fewer type IIs and more type III to type V in FHL compared to the findings of previous studies [
1,
2]. The insertion of the FHL muscle into each toe naturally affects the big toe, but it also affects the second and the more peroneal toes.
The interconnections between the FHL and FDL tendons are well known. The FHL tendon separates into two parts, and one joins with the FDL tendons. This tendinous crossover has been examined previously and is known as “the Knot of Henry” [
13] or “Junctura Tendinum” [
14]. The Knot of Henry is can be found 1.8 cm below the navicular tuberosity and 5.9 cm distal to the medial malleolus [
15]. Although the above ratios varied in individual studies, the majority reported a pattern of tendon binding from FHL to FDL. In our previous studies [
1,
2], the dissection of six cadavers (twelve legs) showed that the FHL tendon bifurcated and joins the FDL in all of them, while there was no tendon coupling from the FDL to the FHL (
Figure 2a–d). In one previous study examining the cadavers’ Knots of Henry in 16 limbs [
13], contraction of the FHL in 11 limbs resulted in not just the big toe motion but the second toes as well as the toes on peroneal side, while contraction of the FDL induced no flexion of the big toe. They also mentioned that the transmission of the tension was unidirectional and was always from the FHL to FDL. In a similar study involving 24 limbs of 24 cadavers [
16], 10 limbs showed that the FHL tendon branched into the FDL (42%). In another study, the dissection of the Knot of Henry in 20 limbs of 10 cadavers [
15] showed that the tendon overlap was unidirectional, from the FHL to the FDL in 15 limbs (75%). Another study [
17] reported a unidirectional tendon from FHL to FDL in all 50 cadavers. Furthermore, a recent study [
18] found a unidirectional tendon from FHL to FDL in 97% of 100 lower limbs in 55 cadavers. Thus, we could assume that FHL greatly affects the movement of the second and/or the third toes [
19].
On the other hand, in not a few cases, the FDL does not move the second toe [
1,
2]. We speculate that this is due to maldevelopment and follows the pattern of the Knot of Henry described above. In these cases, it likely represents the progression of FHL muscle dominance on movements of the second and third toes, with simultaneous and relative regression of the FDL muscle dominance. In fact, our results showed a significant correlation between the NSM of the FHL muscle and the NSM of the FDL muscle, adding support to our speculation.
The present study has several limitations. First, the assessment of CFD before BoNT injection did not include a detailed evaluation of mAS [
20] of the toes or objective confirmation of the appearance of CFD by visual inspection during movement. Evaluation of the treatment outcome was limited to face-to-face interviews of subjective symptoms by the therapist with the patient; only subjective symptoms, such as pain at rest or during transfer or walking, were evaluated. Second, our study was retrospective in nature, and the main goal of each BoNT treatment was not to improve the CFD, and it is possible that sufficient doses of BoNT were not selected for treatment of the CFD. Naturally, each toe was not targeted separately during injection either. Third, the study subjects included those with BoNT injections into the flexor digitorum brevis muscle but not those who got orthosis therapy after BoNT treatment for CFD. Fourth, while the treatment effects on the CFD was examined, we did not take into account the possible contribution of tenodesis on the ankle position to the CFD. Many patients were in position with plantar flexion and ankle inversion, for whom further BoNT was injected simultaneously into the triceps surae and tibialis posterior muscles. The fifth limitation is the varying timing of assessment. The assessment time of the treatment outcome varied from two to six weeks, and thus we may have missed clinical changes in some patients. These points are issues that need to be considered in future studies.
5. Materials and Methods
The protocol of this study was approved by the Ethics Review Committee of the Jikei University School of Medicine for Biomedical research [#26-377(7883)]. Each patient had signed the consent form for BoNT administration before the injection. The study was carried out in compliance with the Helsinki Declaration.
The inclusion criteria were as follows: (1) more than 12 months of history of stroke-related lower limb paralysis its spasticity graded 1 or more on the modified Ashworth scale (mAS) [
20]; (2) Previous diagnosis of “CFD due to spasticity” at the Outpatient Department; (3) Age > 20 years; (4) time between the onset of stroke and current treatment of ≥three months; (5) No contraindication to BoNT administration [
21,
22]; and (6) Previous BoNT injection to the FHL or FDL muscles, along with the electrical stimulation of CFD, between 1 August 2013 and 28 February 2021. The data of patients who opted out of the study were excluded from analysis.
The medical history of each subject was collected from the following six facilities in Japan: The Jikei University Kashiwa Hospital (Kashiwa, Chiba, Japan), The Jikei University Daisan Hospital (Komae, Tokyo, Japan), Tokyo Teishin Hospital (Tokyo, Japan), Medical Center Narita Hospital (Narita, Chiba, Japan), Kita-Kashiwa Rehabilitation General Hospital (Kashiwa, Chiba, Japan), and the Motoyama Rehabilitation Hospital (Kobe, Hyogo, Japan).
The analyzed patients were those for whom we performed BoNT injection into the spastic muscles for the upper/lower extremity of the paralytic side (
Figure 1). Before administration, we diluted BoNT with saline down to 1.25–2.5 units/0.1 mL, then injected with neuromuscular electoral stimulation as guidance. We used 25-gauge sterile pole anesthesia needles (Top Co., Tokyo, Japan) for FHL and FDL muscle injections. The number of BoNT units injected into the FHL, FDL, flexor digitorum brevis (FDB), and quadratus plantae (QP) muscles was estimated from the medical records.
We applied the electoral stimulation using a New Tracer NT-11 (Top Co., Tokyo, Japan). The sites and directions of the needle insertion are shown as in
Figure 3. For each patient, we confirmed that the stimulation needle was in the right place (i.e., FHL or FDL muscles) by virtually recognizing the muscle contraction/toe movement during the stimulation. A precaution was taken not to inject the BoNT to the tibial nerve, located close to the FHL muscle. Because it would create a painful muscle contraction, we judged that the tibial nerve was stimulated when/if the entire foot contracted, unlike when the FHL or FDL muscle was focally stimulated. Two physicians or a physician and a nurse observed and recorded the effects of FHL/FDL muscle contraction on toe movements.
In the next step, we evaluated the number of toes controlled by the contraction of FHL and FDL based on differences in the response to electrical stimulation. We retrospectively sorted the muscle contractions by their strength: + (marked muscle contraction), ± (weak contraction), and − (no contraction), as described previously [
1,
2]. Furthermore, we also analyzed the number of toes that moved with the electrical stimulation of the FHL muscle and classified them. Type I represented movement of the first toe only upon FHL stimulation, type II was for the first and second toes, type III was for the first to third toes, and type IV was for the first to fourth toes. Furthermore, we used “>” if the contraction was weak on the peroneal side of the stimulated toes, and put no sign if it was strong. Similarly, for the FDL muscle, we classified them into type I as those with movement of the first to fifth toes, type II of the second to fifth toes, type III of the third to fifth toes, and type IV of the fourth and fifth toes only. Similar to the FHL muscle, we put no sign if a strong contraction on the most tibial side of the toes and the fifth toe was recorded. We put “<” for weak contraction in the most tibial side of the toes, and weak contraction of the fifth toe, i.e., the most peroneal side, was marked as “>” [
2].
We used the median value in patients who underwent repeated tests described above on the control of toe movement. To calculate the median value, the numbers were placed in the following order: I, >II, II, >III, III, >IV, IV, and >V for FHL; and I, <I, >II, II, <>II, <II, III, <>III, <III, IV, <>IV, <IV, and V for FDL. When the median value could not be determined, the most frequent value was used in evaluation of the state of control for the patient in question, and when even that could not be determined, we used the number of the first evaluation to determine the state of control. Based on these procedures, we determined the state of control (qualitative variables and ordinal scales) for each toe of the FHL and FDL muscles for all patients (MCT: the mode of control over the toes). However, if one of the two muscles was never stimulated, it was not possible to determine the state of control of that muscle over the toe in that patient.
In the next step, we assessed the degree of “variability” in FHL and FDL control of each toe using the data of patients who underwent several toe muscle stimulation tests. In this procedure, weak movements of the toes were ignored, and the number of strong toe movements was used as a quantitative variable (NSM: number of strong toe movements). The standard deviation of the NSM (SD-NSM) was used to assess the “variability” within the same subject. For FHL and FDL muscles, we calculated the mean value and standard deviation of SD-NSM using the data of all subjects who underwent repeated electrical stimulation tests.
The medical records of the participants were searched for improvements or worsening of clinical signs and symptoms of CFD at 14–42 days after BoNT injection. The medical records were written based on the attending physicians’ clinical assessment as well as on the direct interview with the patients/relatives. The assessor was blinded to the results of FHL and FDL electrical stimulation and the underlying clinical condition of the patient. Specifically, we collected data on post-treatment subjective symptoms (124 BoNT injections in 53 patients) from the 146 data points (
Figure 1). Next, in the 124 BoNT injections (with treatment outcome documented in the medical records), we compared and examined the relationship between MCT and NSM and the effects of BoNT injections into the FHL and FDL muscles, respectively, including the effects in those patients who received several injections. Furthermore, analysis of the FDL muscle included an adjustment for factors (underlying conditions) that may have an impact on the objective variable in order to remove confounding (
Table 3).
We then examined the relationship between MCT and NSM and the efficacy of the first single BoNT injection in 53 patients (
Figure 1) after adjustment for factors related to underlying conditions that could affect the objective variable in particular in order to remove confounding (
Table 4). Finally, for the same patients, we examined the correlation between the NSM of FHL and FDL muscles by simple regression analysis.
Baseline subject characteristics are presented as frequencies and proportions for categorical data, and summary statistics (number of subjects, mean ± standard deviation, or median with interquartile range) are presented for continuous data. Logistic regression analysis was used to compare two groups of patients with and without subjective efficacy of initial treatment for the MCT and NSM of the two muscles. The treatment outcome was the objective variable (validity: 1, invalid: 0), and the MCT or NSM of the FHL or FDL muscle and the causative disease (cerebral hemorrhage: 1, cerebral infarction: 0) were the explanatory variables. The generalized estimating equations (GEEs) for the logistic regression model were used to test for all repeated BoNT injections in the same patient, including the second and subsequent injections. The presence or absence of treatment effect (validity: 1, invalid: 0) was used as the objective variable. The model was adjusted for the MCT and NSM for FHL or FDL muscle and the underlying condition (cerebral hemorrhage: 1, cerebral infarction: 0). Pearson’s correlation coefficient was used to examine the correlation between the NSM of the FHL muscle and the NSM of the FDL muscle at the time of the first injection. Two-tailed tests were performed and p values of <0.05 denoted the presence of statistically significant differences. All analyses were performed using the SPSS statistics software (ver. 26, IBM Japan, Tokyo).