Spasticity after stroke is a common sensorimotor control disorder, with an incidence rate of 17 to 42.6% being observed among stroke patients (1). If spasticity is not promptly and effectively treated, it can lead to soft tissue contractures, thereby potentially resulting in the complete loss of motor function. Spasticity, whether occurring alone or with contracture, not only compromises motor recovery after stroke but also induces pain, heightens caregiver burden, and increases the risk of depression—consequently prolonging rehabilitation periods and substantially raising medical costs (2–4). Therefore, the prevention and treatment of spasticity/contracture after stroke has become an urgent problem in the field of rehabilitation medicine.

The standard clinical assessment for spasticity and contracture involves passively stretching the affected muscle to determine the onset and quality of resistance (5, 6). Both conditions present with increased resistance to passive movement. However, accumulating evidence indicates that the Modified Ashworth Scale (MAS) cannot differentiate between these two distinct pathological entities (3, 7). Consequently, in clinical practice they are frequently confused or described by interchangeable terms such as hypertonia, stiffness, or hypoextensibility (7, 8). Therefore, the European consensus has standardised the terminology for this phenomenon, adoptingthe term “hyper-resistance” to describe impaired neuromuscular response during passive stretching (9). In this study, “spasticity” is defined according to Lance (10) as a motor disorder characterised by a velocity-dependent increase in tonic stretch reflexes with exaggerated deep tendon reflexes, and as a positive manifestation of the upper motor neuron syndrome. This constitutes the neurogenic component of hyper-resistance (9). “Contracture” refers to increased resistance arising from non-neurogenic factors, including structural changes in soft tissues such as muscles, tendons, fascia, joint capsules, and associated connective tissues (11). This represents the non-neurogenic component of hyper-resistance (9). Therefore, hyper-resistance encompasses both spasticity and contracture, and the two conditions frequently coexist and are readily conflated in clinical practice. Currently, there is no universally agreed diagnostic criterion for contracture in clinical practice, and different studies have adopted varying standards, namely the minimal important change (MIC) and minimal clinically important difference (MCID) in passive range of motion (PROM) across joints (12). Mehrholz and colleagues proposed a 10° change in PROM as the MCID (13), whereas a recent Cochrane systematic review advocated a 10% change relative to the normal PROM as the MCID for contracture assessment (11). In the present study, we adopted this 10% threshold as the criterion for contracture, with the reference range for passive ankle motion being 70° (45° plantarflexion and 25° dorsiflexion), thereby defining contracture as a PROM limitation of ≥7°.

Previous clinical studies have mainly focused on observing the efficacy of BoNT-A in terms of treating spasticity after stroke (14, 15). Although BoNT-A can reduce the tension of spastic muscles, its effects on patients’ motor function, daily living ability, and gait are quite different. One possible reason for these findings is that the included research subjects were not consistent, and most of the studies did not distinguish whether the hyper-resistance of the muscles involved simple spasticity or spasticity accompanied by contracture. Therefore, we designed this retrospective clinical study to observe the clinical efficacy of BoNT-A injections in different types of post-stroke hyper-resistance in the ankle plantar-flexor muscles.

The main finding of this study is that BoNT-A effectively ameliorates the severity of post-stroke hyper-resistance, whilst enhancing motor function and activities of daily living. Significant improvements in MAS and BRS were observed as early as 2 weeks following BoNT-A injection; however, amelioration of FMA-lower extremity and BI was not evident until 4 weeks or 12 weeks. Furthermore, the co-occurrence of contracture with BoNT-A treatment is associated with diminished therapeutic outcomes, particularly regarding MAS and BRS scores.

In the group with both spasticity and contracture, the improvement of MAS by BoNT-A was less than that in the group with simple spasticity, and the optimal therapeutic effect occurred later, at 4 weeks. For patients with both spasticity and contracture, the muscle-tendon units were in a shortened state. At 2 weeks, BoNT-A could have a good blocking effect on the neuromuscular junction, and the resistance from the neurogenic component had improved, but the resistance from the non-neurogenic component had not been well relieved, resulting in less improvement of MAS than the group with simple spasticity. Additionally, after all patients received BoNT-A injection, they all routinely received traditional rehabilitation treatment, which would alleviate the non-neurogenic contribution to increased resistance and thus accounts for the delayed peak response in the spasticity-with-contracture group until week 4. The improvement in lower-limb motor function following stroke with BoNT-A may be attributed to the reduction of increased resistance in the ankle plantar flexors, thereby permitting more pronounced ankle dorsiflexion. This may also explain why the BRS improvement in the spasticity-with-contracture group was less marked than in the spasticity group. The improvement in lower limb FMA was less evident than that of BRS. A plausible explanation is that the patients in our study had relatively preserved motor function, with FMA scores around 20, which may have introduced a ceiling effect and contributed to the lack of statistical significance in the final results. Both groups showed improvement in BI scores. This improvement did not occur immediately following injection but emerged between weeks 4 and 12, suggesting that the amelioration of patients’ activities of daily living was not attributable to the immediate relief of increased muscle resistance, but rather progressed in parallel with the improvement in lower limb motor function. In other words, BoNT-A injections do not directly improve FMA scores or activities of daily living, and the co-occurrence of contractures likewise shows no substantial correspondence with either of these two aspects.

To our knowledge, this is the first retrospective study to directly compare the time course of BoNT-A efficacy between post-stroke patients with spasticity alone and those with co-existing spasticity and contracture using a multimodal assessment. While Gracies et al. previously demonstrated that severe soft tissue shortening impairs active functional goal attainment in a post-hoc analysis of the ULIS-II cohort, their study focused on the upper limb and did not examine the differential temporal dynamics of MAS, motor function, and ADL improvements (16). Our findings extend these observations to the lower limb and reveal a previously unreported dissociation between the time courses of resistance reduction, motor recovery, and functional independence. Consistent with Gracies et al., we found that the presence of contracture attenuates the magnitude of MAS improvement following BoNT-A injection (16). However, our study further demonstrates that this attenuation is evident as early as week 2 and persists throughout the 12-week observation period. Lindsay et al. demonstrated in a randomised controlled trial that early BoNT-A intervention (mean 18 days post-stroke) significantly reduces contracture development compared with delayed treatment; however, their design precluded direct comparison of established contracture versus spasticity alone, and they did not assess the temporal dissociation between spasticity reduction and functional gains (17). Falcone et al. (18), in a 14-year retrospective cohort, observed that chronic-phase patients required escalating BoNT-A doses, suggesting diminished treatment responsiveness over time, but they did not stratify outcomes by contracture status or report discrete time-point analyses. These findings are further contextualised by recent evidence that conventional rehabilitation, particularly muscle stretching, can mitigate non-neurogenic resistance and enhance BoNT-A long-term efficacy (19), and that multimodal approaches combining BoNT-A with adjunctive therapies such as extracorporeal shock wave therapy may be required for optimal management of contracture-plus-spasticity (20). The 2026 AHA/ASA scientific statement on post-stroke spasticity emphasises the necessity of distinguishing neurogenic from non-neurogenic contributions to increased resistance (3), underscoring the clinical relevance of our stratified approach. Future prospective studies with standardised rehabilitation protocols and longer follow-up periods are warranted to validate these retrospective observations and to optimise treatment algorithms for this challenging patient population.

This study has several limitations. First, the retrospective design precludes causal inference. Selection bias may exist in the allocation of patients to groups, and the observed associations may reflect confounding by disease chronicity, severity, or unmeasured factors. Additionally, baseline characteristics were not balanced between groups, as patients in the spasticity-with-contracture group had longer disease duration than those in the spasticity-only group, which likely confounded the observed treatment responses. The between-group difference in disease duration precludes attributing the delayed treatment response solely to contracture presence; chronicity itself may independently attenuate BoNT-A efficacy. Second, rehabilitation protocols were not standardised across the cohort, which may have further influenced functional outcomes. Third, the sample size was relatively small, limiting the generalisability of our findings. We did not perform multivariate adjustment due to sample size constraints; future prospective studies with larger cohorts are warranted to adjust for disease duration, baseline severity, and rehabilitation intensity. Fourth, the follow-up period was confined to 12 weeks; longer observation is required to ascertain the durability of BoNT-A effects in patients with established contracture. Finally, contracture severity was dichotomised rather than graded, potentially obscuring dose–response relationships. Future prospective, multicentre studies with standardised concomitant therapies, balanced baseline characteristics, and extended follow-up are warranted to validate these observations.

BoNT-A markedly reduces post-stroke hyper-resistance and enhances motor function and activities of daily living; by contrast, concomitant contracture is associated with delayed and attenuated improvement in MAS and BRS.

Based on our findings, we recommend early BoNT-A injection for post-stroke spasticity to prevent contracture development and maximize therapeutic benefit. For post-stroke muscle hyper-resistance, a precise assessment prior to treatment appears necessary to guide subsequent therapeutic interventions, particularly botulinum toxin type A injections. Recently, we have also been exploring a multimodal assessment approach incorporating surface electromyography, isokinetic dynamometry, two-dimensional muscle ultrasound parameters, and elastography, in order to identify the pathological origins of muscle hyper-resistance. BoNT-A should be combined with complementary rehabilitation measures, including soft tissue stretching and joint mobilization, rather than used as monotherapy.