Introduction

Spasticity is a type of hypertonia resulting from damage to the central nervous system and indicative of upper motor neuron dysfunction [1]. Traditional, physiological definitions describe spasticity as “a motor disorder characterized by a velocity-dependent increase in tonic stretch reflexes (muscle tone) with exaggerated tendon jerks, resulting from hyperexcitability of the stretch reflex” [2]. However, there has been a historical lack of consensus regarding the definition and measurement of spasticity, which can manifest in a number of ways besides abnormal stretch reflex behaviour, including pronounced muscle stiffness and involuntary spasms. In lieu of a unified description [3], contemporary definitions attempt to balance mechanistic precision and clinical applicability, defining spasticity broadly as “disordered sensorimotor control, resulting from an upper motor neuron lesion, presenting as intermittent or sustained involuntary activation of muscles” [4]. Clinically, spasticity can lead to reduced contractibility, a heightened risk of pain, contractures, and changes in muscular structure and control that compromise mobility and positioning. The incidence of spasticity in spinal cord injury (SCI) is estimated to range broadly from approximately 65 [5] to 93% [6], though it is frequently underdiagnosed in SCI populations [7] on account of its variable clinical presentation and time-dependent progression [8]. While some people with SCI find practical benefits to spasticity, such as aiding transfers and blood pressure regulation [7], spasticity can more commonly have considerable negative effects [9] relating to pain [10], functional performance [11], and mental wellbeing [12].

Passive movement (PM) techniques are here defined as movements of a person’s joints performed entirely by an external force without active muscle contraction [13]. PM can be delivered manually [14] or mechanically [15], ranging from traditional stretching exercises to automated continuous passive motion (CPM) devices. PM techniques are proposed to attenuate spasticity [13, 16] through neural mechanisms (e.g., sensory input modulation, induced reflex hyperexcitability) and non-neural mechanisms (e.g., adaptations in soft tissues) [17], though long-term improvements are contested. The SCI Physiotherapy Guidelines weakly recommend against the use of passive range of motion (PROM) interventions to treat spasticity in people with SCI [18] due to the absence of high-quality studies and meta-analyses, an assessment of benefits and harms, personal experience, and other relevant considerations. This highlights the inconsistencies between recommendation and practice which serve as motivation for this review. While PROM represents one distinctive form of PM, other modalities, such as CPM, passive stretching, and joint mobilization are not covered by this recommendation. In the wider PM literature, existing reviews either broadly summarized physiotherapeutic modalities for individuals with SCI [19], or focused on the effects of passive cycling in SCI groups [20, 21] without a singular focus on spasticity as an outcome measure. Consequently, this scoping review aims to systematically search and synthesize all existing evidence concerning the use of PM interventions targeting spasticity outcomes in individuals with SCI during rehabilitation.

Methods

This scoping review is reported in alignment with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews (PRISMA-ScR) [22] checklist and explanation document. The study protocol was pre-registered in Open Science Framework on 26th November 2024 (DOI 10.17605/OSF.IO/JPCV3).

Information sources and search strategy

The reporting of the search strategy performed in this scoping review conforms to the PRISMA Statement for Reporting Literature Searches in Systematic Reviews (PRISMA-S) [23]. Two search blocks were designed to compile search terms closely associated with i) spinal cord injuries and ii) passive range of motion interventions. The search terms were determined in consultation with clinical experts and via a thorough reading of relevant scientific literature. Seven databases were chosen to run the search, specifically: Embase (via OVID), Medline (via OVID), PsycInfo (via OVID), Scopus, Web of Science, CINAHL (via EBSCO), and CENTRAL. No filters were applied to restrict languages, study design, population characteristics, or source type. Each database was searched from its earliest available records to its most recent, as of 25th November 2024. (Medical) subject headings were incorporated and adjusted depending on the indexing of each database. Upon the addition of a new (medical) subject heading, a corresponding free-text search term was introduced to the search string. Search terms were entered consistently across the seven databases. Searching within blocks was facilitated by the Boolean Operator “OR”. Searching across blocks was facilitated by the Boolean Operator “AND”. The search string was piloted multiple times prior to performing the final searches. The final search string is provided in Supplement 1.

Additional searches

As a supplement to the database searches described above, a number of additional searches were conducted by a trained student assistant to identify any relevant studies potentially missed through the systematic search process. In Google Scholar [24], search terms from the population block (e.g., “spinal cord lesion*”) and the intervention block (e.g., “passive range of motion*”), were combined and searched together. On 26th February 2025, five such combinations were constructed, and the first 100 results of each search sorted by relevance were screened at the title and abstract level. On the same date, the search term “spinal cord injury” was entered into PEDro [25] to assess the relevance of all retrieved clinical trials. In an over-inclusive manner, potentially relevant studies found via Google Scholar and PEDro were imported into Covidence to facilitate independent double-screening at the full-text level.

Additionally, all included studies were subject to forward and backward snowballing [26]. This also applied to ‘tagged’ studies which contained relevant content but had to be excluded because of their format. Forward snowballing, designed to identify new materials based on all subsequent citations of a given article, was achieved using Google Scholar. Backward snowballing, designed to identify new materials based on all cited materials within a given article, was achieved manually through access to reference lists.

Prior to data extraction, Google searches were performed with the intention of verifying that no erratums, addendums, or other published information connected to the included studies alters interpretation of their findings.

Eligibility criteria

Studies were required to be in total agreement with the protocol’s predefined eligibility criteria to be included in the present review. Namely, studies had to be written in English, Danish, Norwegian, or Swedish, reflecting the linguistic competencies of the author group, and to have reported original data from a peer-reviewed academic journal as a full-text article. SCI participants had to constitute ≥50% of the study population, or their data needed to be reported separately. Per our inclusive definition, both static and repetitive PM interventions were eligible for inclusion. Studies were not required to explicitly refer to their intervention as PM, as long as treatment consisted solely of passive movement of the joint. PM techniques could be either manual or mechanical and were required to be performed in a rehabilitation context, defined as structured therapeutic interventions delivered by healthcare professionals with a rehabilitative aim. This necessarily excludes non-therapeutic interventions and diagnostic studies. The intervention must have consisted solely of PM techniques, thereby necessarily excluding PM techniques combined with treatment such as strength training, electrical stimulation, or active exercise. Given the observed heterogeneity of outcome measures and lack of consensus on how best to operationalize spasticity, the impact of the intervention needed to be evaluated using any spasticity outcome falling within Pandyan et al.’s definition [4], such as H-reflex amplitude or Modified Ashworth Scale (MAS) scores. No restrictions were imposed on the duration of outcome assessment, allowing for the inclusion of immediate, short-, mid-, and long-term effects.

Screening procedure

All records were imported into the systematic review software Covidence [27], where duplicates were automatically removed. During screening, all remaining duplicates were removed manually.

Screening guides were prepared by SDW to ensure a consistent and reproducible process from the beginning. To further ensure alignment in the screening process, independent double screening of the first 100 records was piloted between SDW, AOE, SLR, and a trained student assistant. Each record underwent double screening, meaning that two reviewers were required to independently screen each record before it was included or excluded. Based on this piloting, initial differences in interpretation of the eligibility criteria were identified and discussed within the screening group to foster a unified approach throughout the screening phase. Following the pilot phase, independent double-screening at the title and abstract level proceeded between SDW, AOE, SLR, and the trained student assistant. Disagreements during screening were discussed and resolved internally. All studies forwarded from the title and abstract screening level were then independently double-screened at the full-text level by SDW, AOE, and the trained student assistant. As before, disagreements pertaining to individual studies were discussed and resolved within the screening group, and via consultation with MRV where appropriate. Reasons for exclusion at the full-text level were documented for reference. Records containing relevant content but published as reviews, book chapters, protocols, and other ineligible formats as per the eligibility criteria received a tag in Covidence. Tagged studies constituted the foundation of our additional searches, described below. Information missing from the full-text but considered essential to the decision making process was sought from corresponding authors. Records independently double-screened forward from the full-text screening level were extracted (described below) and included in the review.

Data extraction

Relevant data were independently extracted by SDW and AOE into the same, pre-defined categories, housed within the same, standardized publication table, produced in Microsoft Word. Data considered relevant to extract comprised of: author group, year of publication, country of origin, sample size (n), sample characteristics (age [range and mean/median], sex [percentages]), injury characteristics (injury level, injury completeness, and time since injury), setting, study design, intervention type, relevant outcome(s), and relevant main results. An additional column was used to collect all information provided by corresponding authors. Discrepancies encountered during data extraction were resolved between SDW, AOE, and, where appropriate, MRV, by jointly consulting the article in question and reaching a consensus. The full extraction table is provided as Supplements 2 and 3.

Synthesis methods

Data were narratively synthesized using the extraction table and according to the SWiM reporting guidelines [28]. Included studies were grouped by intervention type, and outcomes were extracted as reported. No studies were featured more prominently than others in the reporting with respect to study quality. Heterogeneity in the findings was determined, including duration and frequency of intervention, and movement speed. Limitations of the evidence base were noted for discussion in lieu of a formal risk-of-bias or GRADE assessment [29]. The stepwise screening process was visualized in a PRISMA-compliant flow diagram [30] (Fig. 1).

Fig. 1
figure 1

PRISMA-compliant flow diagram illustrating study selection.

Full size image

Results

Study flow

Figure 1 illustrates study flow. In summary, a total of 3402 records were retrieved across the seven databases. First, 1774 duplicates were removed from the screening process, leaving 1628 unique records to be screened at the title-and-abstract level. A total of 1270 studies were excluded at this level, meaning 358 records remained for full-text screening. Three studies were inaccessible and therefore not screened at the full-text level. Information regarding inaccessible studies is provided as Supplement 4. In assessing the 355 accessible studies, ineligible interventions (n = 127), publication formats (n = 98), outcomes (n = 63), populations (n = 35), and publication languages (n = 19) were registered as main causes for exclusion. One study of potential relevance was identified via supplementary searches but was excluded at the full-text screening level. Therefore, a total of 13 studies were included in the scoping review.

Study characteristics

Included studies were published between 1995 and 2018, with six in the 2000s [31,32,33,34,35,36], five in the 2010s [15, 37,38,39,40], and two in the 1990s [41, 42]. Studies were diverse in terms of research design, including randomized controlled trials [15, 32, 40], repeated measures designs [31, 33, 38], crossover studies [35, 37], observational studies [41, 42], case reports [34, 39], and one pre-post intervention study [36]. Intervention group sample sizes ranged from 1–35, with 5 studies including 1–5 participants, 5 studies including 6–10 participants, and 3 studies including 11 or more participants. Of the 10 studies reporting sex data, more men (n = 62) than women (n = 12) participated in experimental conditions. Excluding case studies of 24-year-old [34] and 67-year-old [39] men, mean ages of SCI intervention group participants ranged from 30.1 [38] to 46.6 [35] years old. Injury characteristics were reported heterogeneously, though ASIA levels ranged from C2 to L11 and mean time since injury ranged from 13 months [34] to 13 years [36].

Spasticity was measured diversely using clinical scales [15, 31, 32, 35, 36, 39, 40], neurophysiological measures [31, 34, 38, 40,41,42], functional tests [33, 35, 37, 39], and subjective reports [32, 33]. The most common outcome measures were H-reflex amplitude or modulation (indicative of spinal reflex excitability) [15, 31, 34, 38, 40,41,42] and MAS scores, used to quantify spasticity severity [15, 31, 32, 35, 36, 40]. Seven studies reported outcomes resulting from a single PM session [31, 33, 35, 37, 38, 41, 42], while the number of sessions across the remaining studies ranged from 12 [36] to approximately 168 [40].

Narrative synthesis

The results presented below were organized into three categories according to type of PM application, namely, i) Passive cycling and robotic PM, ii) CPM, and iii) manual PROM. Table 1 summarizes the study characteristics and key findings of the included studies.

Table 1 Summary of the included studies, containing author(s), study design, spasticity measure(s), intervention characteristics, and key findings.
Full size table

Category 1: passive cycling and robotic PM

A total of eight studies evaluated passive cycling and robotic PM interventions [33,34,35, 38,39,40,41,42].

In one longitudinal RCT study, 35 participants received thrice-daily electrical passive cycling (EPPS) for two months [40]. Compared to a control group, optimal-level EPPS significantly decreased MAS scores (p = 0.003) one-year post-intervention [40]. While significant bilateral reductions in Hmax/Mmax ratios from 0.40–0.20 (p < 0.001) were recorded, H-reflex responses remained unchanged [40].

Four studies reported short-term improvements from a single PM trial [35, 38, 41, 42]. In three tetraplegic participants, passive single-leg pedalling at 10 rpm resulted in a significant mean reduction in H-reflex amplitude of 35% in the active leg and 45% in the stationary, contralateral leg (p = ≤0.05) [41]. Comparatively, the ankles of four participants passively moved by a servo-controlled electromechanical actuator resulted in sustained H-reflex depressions, lasting from 550 ms–4000 ms after onset of ramp-and-hold dorsi-flexions [42]. Alternative movement patterns, such as single and double dorsiflexion (300–1300 ms) and continuous dorsiflexion (approx. 150 ms), showed weaker reductions [42]. Elsewhere, robot-assisted passive ankle movements performed at high-speed (50 cycles/min) and low-speed (20 cycles/min) for eight minutes in ten participants were found to reduce H-reflex amplitudes for up to 20 min post-intervention, with effects observed sooner (within 10 min) for high-speed passive ankle movement [38]. Improvements in muscle resistance were speed-dependent, insofar as low-speed passive ankle movement reduced resistance during slow stretches, and high-speed passive ankle movement reduced resistance during fast stretches [38]. Five passive leg cycling participants demonstrated significant average reductions in the MAS for both left and right legs (p < 0.05) [35]. The relaxation index increased by approximately 12% on average post-treatment, with a significant improvement observed in the left leg (p < 0.05) [35].

Two case reports demonstrated short-term benefits of passive cycling interventions on spasticity [34, 39]. Following a 13-week motorized bicycle exercise trainer programme, the H-reflex amplitudes of a 24-year-old male (ASIA B, C7) were significantly reduced from 76–27% at 5 Hz and from 65—12% at 10 Hz by the 12-week assessment [34]. However, four weeks post-intervention, H-reflex amplitudes regressed (57% at 5 Hz and 52% at 10 Hz), and the participant reported that his spasticity had returned two weeks prior [34]. Similarly, a 67-year-old male (ASIA A, T4 SCI) performed thrice-weekly passive cycling at a maximum of 30 rpm, lasting five weeks [39]. Initial knee flexion velocity during the pendulum test (F1-VEL) was significantly increased by the end of testing and up to two weeks later [39]. While spasticity returned to pre-intervention levels by 3-weeks post-intervention [39], Spinal Cord Injury Spasticity Evaluation Tool scores suggested a lasting reduction in the impact of spasticity on daily life [39].

In the eighth and final study in this category, a group of 10 motor-complete SCI participants performed a single 30 min session of passive cycling at 40 rpm, resulting in no significant improvement in peak torque outcomes [33]. While six participants self-reported feeling less spastic after the intervention, three participants also reported reductions after a static control condition [33].

Category 2: continuous passive motion (CPM)

Four studies evaluated CPM interventions [15, 31, 36, 37]. In a pre-post study, 12 participants received twice-weekly 30 min passive leg movements using motorized tables for six weeks [36]. A 50% immediate reduction in MAS and a 90% reduction in VAS were observed [36]. Participants who adhered to the six-week protocol experienced a significant (p < 0.018) 30% reduction in self-reported spasticity that persisted for at least one-week post-treatment [36]. For eight SCI participants, sixty minutes of ankle CPM significantly reduced soleus H-reflex amplitude to 77 ± 32% of its initial value immediately after treatment (p = 0.0476) but regressed to 93 ± 36% after 10 min [31]. A sham group showed no change in H-reflex amplitude [31]. Despite partial recovery in H-reflex amplitude, participants registered significant improvements in MAS scores for 10 min post-treatment (p = 0.016; Z = –2.414) [31]. Additionally, four consecutive weeks of hour-long daily ankle CPM resulted in significant average decreases in MAS scores (p = 0.013) and improved post-activation depression (PAD) of the H-reflex (F₁ = 7; p = 0.038) in a group of seven RCT participants [15]. One participant who completed an additional eight weeks of ankle CPM training maintained gains but experienced diminishing returns after the initial four-week protocol [15]. Further, a single thirty-minute CPM session significantly increased first swing excursion (p = < 0.05) with a large immediate effect size (Cohen’s d = 1.27) for 10 participants when compared to sham-controls [37]. Unlike manual stretching, CPM-related improvements to first swing excursion persisted for up to 45 min post-treatment (Cohen’s d = 1.23) [37]. However, within-group changes were not statistically significant (p = 0.085 for immediate, p = 0.260 for delayed) [37]. While five participants demonstrated a clinically meaningful reduction in spasticity (≥12° increase in first swing excursion), and three participants demonstrated moderate reductions (6–11° increase), two participants experienced no benefit [37].

Category 3: manual prom

Only one study evaluated a manual PROM intervention, in which a series of twice daily, 10 min passive ankle movements were provided to 20 participants [32]. Only one ankle received treatment, which was performed for five days a week over a period of six months [32]. While participants subjectively reported reduced spasticity, between-group differences in MAS scores for the treated and untreated ankles lacked statistical significance, as indicated by 95% confidence intervals that included zero for both hamstring and plantar-flexor muscles [32].

Discussion

This scoping review mapped and synthesized the available, empirical evidence on PM interventions for managing spasticity outcomes in individuals with SCI during rehabilitation. While not necessarily indicative of clinical relevance, mechanical modalities such as CPM, passive cycling, and robotic PM were consistently associated with reduced spasticity using validated measures including H-reflex amplitude and MAS scores. The few multiple-session, longitudinal PM interventions included in this review demonstrated positive and sustained effects on spasticity during a 12-week training period [15], and at one-week [36], and one-year [40] follow-ups. However, the degree of improvement depended on several methodological factors, such as movement patterns (e.g., single, double, or continuous dorsi–plantar flexion), speed (e.g., rpm), and dosage (e.g., duration and frequency). Elsewhere, improvements were often temporary, returning towards baseline levels within seconds [42], minutes [31, 37, 38], or weeks [34, 39] of the cessation of PM interventions. While this is partially explained by the preponderance of single-session designs with short-follow ups [31, 33, 35, 37, 38, 41, 42], improvements observed at longer-term follow-ups also tended to plateau [15, 39], thereby limiting the ability to infer long-term clinical relevance. Evidence to support the use of manual PROM interventions in reducing spasticity was limited to subjective reports and non-significant MAS scores, which makes it difficult to draw firm conclusions about therapeutic value. To this point, it is important to consider that patient-reported improvements in spasticity were observed even when objective physiological measures showed comparatively limited changes [32, 33, 36, 37]. In light of the competing definitions and measurement tools that complicate its diagnosis and treatment, these findings underscore the need for future research to measure the multifaceted experience of spasticity holistically. Taken together, these observations limit the ability to recommend manual or mechanistic PM interventions as a standalone treatment for spasticity-related issues in an ongoing rehabilitation context.

Study findings in context

These findings are contextualized by those reported in the existing literature. One systematic review from 2021 took a broader approach, assembling all multiple-session physiotherapeutic interventions for individuals with SCI treated for spasticity [19]. In a subchapter dedicated to passive interventions, one CPM intervention [15] and one manual PM intervention [32], both of which were included in the current review, were presented. The subchapter also contained a passive standing exercise intervention [43] and a kinesiotaping intervention [44], both of which lacked the passive movement of joints through their range of motion required for inclusion in this review. Two further reviews have looked closely at passive cycling interventions for individuals with SCI [20, 21]. Nardone et al. [21] reported the effects of passive cycling across two SCI studies [33, 45], concluding that objective reductions in spasticity were limited, though subjective improvements were reported. Our review, with its expanded and updated search strategy and wider inclusion of passive movement modalities, identified both more studies and a greater variability in results. In comparison, Phadke et al. [20] included a more recent and expansive search and found 11 passive leg cycling studies within the SCI literature. Three of the included studies measured spasticity outcomes, including an RCT [40], a crossover trial [35], and a repeated measures design [33], all of which are included in this review. The authors concluded that multiple leg cycling sessions can have neurological benefits, but that more RCTs are required to add greater statistical power and establish clinical significance [20]. Our review supports the view that further research is needed to establish potential neurological effects, and advocates for further investigation into the observed gap between objective physiological changes and subjective patient experiences.

Looking into the broader neurological literature, the reported outcomes of PM interventions on spasticity are scarcely reported. For example, a systematic review concerning the effectiveness of physiotherapy interventions on spasticity in multiple sclerosis included no PM interventions [46]. Likewise, a mini-review of passive exercise training for individuals with hemiplegic stroke [38] found no examples of pure PM interventions modulating spasticity [47]. In addition, a recent umbrella review of physical therapy interventions for post-stroke spasticity found no examples of interventions consisting solely of PM [48]. However, we identified a pre-post study including 12 chronic stroke patients which demonstrated significant, immediate reductions in MAS and VAS scores after 15 min of dynamic-repeated passive ankle motion exercise [49]. This mirrors the potential short-term benefits found in some studies in individuals with SCI.

Limitations of the included studies

While no formal risk of bias assessment was performed on the included studies, there are some important factors to consider. Primarily, the clinical implications of the reported findings are impeded by the scarcity of multiple-session and longitudinal RCTs. More specifically, ten out of thirteen studies included ten or fewer participants with SCI in experimental conditions [15, 31, 33,34,35, 37,38,39, 41, 42], two of which were case studies [34, 39], which significantly limits generalizability and clinical applicability. While one RCT reported 95% confidence intervals [32], all three RCTs relied primarily on p values to convey between-group differences, and none provided standardized effect sizes [15, 32, 40]. Several studies reported incomplete PM parameters, such as range of motion (i.e., the degree of flexion/extension) [39, 41], movement speed or frequency [32, 35, 37], and movement pattern [34, 41]. The absence of such information complicates cross-study comparisons and challenges reproducibility. Finally, several studies failed to report information about assessor [36, 40] and participant [39] blinding of subjective outcome measures, thereby introducing a potential risk of measurement bias.

Limitations of this review

The term ‘spasticity’ is applied inconsistently throughout the literature and is used to describe several related phenomena, such as clonus, muscle spasms, or hypertonia. Ambiguities surrounding the definition of spasticity are compounded by the variety of clinical, electrophysiological, and subjective measures used to assess it. There exists a lack of consensus regarding an optimal measure of muscle overactivity. For instance, while the MAS is commonly used in clinical practice and is therefore chosen to accurately represent the existing literature, some suggest that biomechanical measures like peak torque are more objective indicators of spasticity [41]. This is due to the reported psychometric limitations of the MAS, including poor construct validity, ordinal scaling issues and, crucially, a lack of concept specificity [50]. This heterogeneity complicated the ability to draw comparisons between spasticity outcomes and required us to adopt a broad definition in this review that some may consider overinclusive. Equally, findings from validated measures do not necessarily reflect clinical significance. Since the review did not require that outcomes be explicitly referred to as measurements of spasticity, no strict boundaries were imposed to guide our judgement regarding the relevancy of an outcome. While we remained in consultation with clinical experts during the screening and extraction process, this remains a methodological consideration that obfuscates the findings of the current review.

As per our protocol, this review was exclusively concerned with the therapeutic use of PM techniques. This required the exclusion of mechanistic spasticity assessments without an intention-to-treat principle [51,52,53]. These studies aimed to investigate how PM might modulate spasticity on a neurophysiological basis by isolating certain biological responses. While beyond the scope of this review, we recognize the importance of these studies to help explain the mechanisms underlying PM’s potential role in treating spasticity. The heterogeneity in injury characteristics across the included studies, particularly in terms of time since injury, injury level, and injury completeness, discouraged us from discussing the interaction between PM interventions and spasticity moderated by these clinically meaningful variables. A subgroup analysis organized by injury characteristics may still have provided greater context to the results. Further details about injury characteristics are provided as Supplement 2.

Lastly, our search was limited to studies published in either English, Danish, Norwegian, or Swedish. While assessing studies in additional languages would have required translation tools beyond our resources and have introduced a potential risk of misinterpretation or inconsistency in assessment, we acknowledge that relevant research may be published in other languages.

Clinical implications and avenues for future research

As reflected in the wider SCI exercise and movement literature [54], the small sample sizes and low-level study designs found within our 13 included studies were indicative of a lack of high-quality research in the area. Thus, well-designed RCTs focusing on clinical importance, subjective experience, and long-term outcomes would help to meaningfully establish the interaction between PM interventions and spasticity in SCI rehabilitation contexts. Alternatively, single-case experimental design (SCED) studies may be well-suited to address the inter-individual variability in spasticity response in a notably heterogenous clinical population [55, 56]. SCED studies are able to demonstrate treatment effects with strong internal validity and refine intervention parameters as proof-of-concept data for an eventual RCT study. This, however, can only be achieved on the basis of a clear and consistent definition of spasticity, which should be a central focus of future research. Additionally, therapeutic PM interventions were excluded from the review when combined with pharmacological [57] or physiotherapeutic [58] components in the same course of treatment. However, PM is used in clinical practice in combination with other therapies to offer synergistic or additive benefits. Therefore, an extension to our review might explore whether PM is more effective when used as an adjunct, rather than a standalone intervention, to treat spasticity. As previously intimated, while single-session designs are inherently incapable of demonstrating long-term effects, efforts to scale-up interventions should bear clinical applicability in mind. For instance, Harvey et al. [32] acknowledge that the intensity of their 20 min PROM sessions to single or multiple joints would hinder long-term application. This naturally calls into question the transferability of Rayegani et al.’s [40] protocol of 20 min sets of EPPS repeated three times daily over the course of two months. Therefore, future research into multiple-session PM interventions on spasticity should question the optimal intensity and dosage required to achieve meaningful outcomes in clinical practice. Moreover, further research is needed to establish additional PM parameters, such as frequency, duration, and speed, to maintain therapeutic benefits over time, as well as explore and discuss potential short-, mid-, and long-term effects of PM to guide clinical application. As such, while mechanical PM interventions were consistently associated with short-term reductions in spasticity across several validated neurophysiological and clinical rating scales, it is important to consider that this review did not assess study quality and is lacking in high-quality designs. As such, while the included studies report noteworthy findings, significant gaps remain that should be addressed in future research.