In-Hospital LSVT BIG Training Versus Structured Rehabilitation Treatment in Parkinson’s Disease: Feasibility and Primary Evaluation on Functional and Respiratory Outcomes

Featured Application

These findings highlight the crucial role of the FITT principle (frequency, intensity, time, type) in rehabilitation programs. They suggest that specific interventions, such as LSVT BIG, might offer distinct advantages over traditional care, particularly for improving dynamic balance and effort tolerance. In particular, specific elements of respiratory training and vocalization could be integrated earlier into standard rehabilitation for patients with Parkinson’s disease. Whether motivational factors made a difference to patients undergoing LSVT treatment remains to be evaluated.

Abstract

Lee Silverman Voice Treatment (LSVT) BIG, primarily developed for outpatient use, is a prominent intervention for patients with Parkinson’s disease thanks to its high-intensity, repetitive exercises involving large movements. This study first evaluated the feasibility of an in-hospital LSVT BIG training program by assessing recruitment capability, compliance, and adherence. The secondary objective was to evaluate the effects of LSVT BIG training on gait, balance, and functional outcomes, as well as respiratory function and quality of life, in comparison with a progressive structured rehabilitation program (SC) of similar intensity and frequency. In-hospital LSVT BIG training for people with Parkinson’s disease was feasible, with 95% recruitment rates and 100% safety and adherence. SC (n = 19) and LSVT BIG (n = 19) significantly improved (for all, p < 0.05) pre-to-post balance (MiniBESTest) and lower limb effort tolerance (6MWT). Delta changes between groups favored LSVT for upper limb effort tolerance (UULEX level, time, p < 0.001), gait speed, and UULEX SatO₂ mean, PCEF, MiniBESTest and 6MWT (for all, p < 0.05). Evaluation of the probability associated with the LSVT BIG showed MiniBESTest as being 8.5 times more likely to exceed the MCID compared to SC. Quality of life was unchanged across both groups. This study successfully demonstrates the feasibility of in-hospital LSVT-BIG^® training, and comparison of outcomes, although exploratory and underpowered, showed better improvements in mobility, balance, and effort tolerance, suggesting a complementary role within traditional rehabilitation protocols.

1. Introduction

Parkinson’s disease (PD) is a progressive neurodegenerative disorder affecting the basal ganglia. It is characterized by motor symptoms such as resting tremor, rigidity, bradykinesia, hypokinesia, and postural instability. These are often accompanied by non-motor symptoms, including sensory and emotional changes, autonomic abnormalities, and loss of higher cognitive functions [1]. PD affects 1–2 in 1000 of the population at any time and the prevalence of PD increases with age, affecting 1% of people over 60 years old [2]. Although pharmacological and surgical treatments provide symptomatic relief, patients’ motor deficits continue to deteriorate as the disease progresses, even with optimized therapies [3].

Physical exercise (PE) is an essential adjunctive therapy for patients with PD [4]. It helps maintain mobility and movement, and while it is unlikely to influence disease progression, it can improve daily functioning by teaching and training PD patients in the use of compensatory movement strategies [5,6]. Interventions like aerobic exercise have been shown to be relatively safe for this population [7]. Although there is a consensus on the beneficial effects of physiotherapy on motor capacity, specifically gait, and balance control, a consensus is still needed on the most effective treatment and meaningful outcome measures [8]. A rehabilitative program for PD should be ‘goal-based’ and focus on practicing specific activities in core areas. This must take into account crucial practice variables such as intensity, specificity, and complexity [8].

Several types of physical exercise have been proposed as effective for PD patients, including sensory cueing (using visual or auditory cues) [9,10], repetitive training of specific movements [11,12], Nordic walking [13], home training programs [14], and musculoskeletal exercises to improve strength, range of motion, and endurance [15]. While all these approaches have been proven to be effective, there is little evidence to suggest that one intervention is superior to another [7]. The lack of consensus arises from heterogeneity of studies and limited protocol standardization, rather than an absolute absence of differences between interventions.

Several large human clinical trials have been completed and collectively support the use of aerobic exercise—specifically, high-intensity aerobic exercise—in improving PD motor symptoms [16]. Common forms of exercise include walking and stationary cycling. Treadmill training employs the specificity of training principles and facilitates relatively robust improvements in gait, including velocity and stride length [16]. This provides a basis for including aerobic exercise as an integral component in the treatment of PD. Over the past decade, the Lee Silverman Voice Treatment (LSVT) BIG has been a well-known intervention for patients with PD [17]. This intensive physical and occupational therapy program is designed to recalibrate disturbed movement amplitude scaling, a common motor deficit in PD. It features high-intensity, repetitive exercises with increasing complexity to improve the perception and execution of larger movements in the upper and lower limbs [16]. The Berlin study [17] was among the first to show that LSVT BIG could effectively improve motor performance.

However, despite promising initial findings, a recent Cochrane review [7] has raised doubts about the effects of strength/resistance training on movement flexibility by LSVT BIG. This uncertainty stems from the heterogeneity of the studies with a limited number of high-quality randomized controlled trials for PD [18,19,20] that have compared LSVT BIG against active control exercises. Existing studies often suffer from methodological limitations, such as a lack of randomization, small patient numbers (many are case series), and variability in settings (mainly home-based and community-based settings and outpatient clinic) [7]. To date, there is a gap in research comparing different training modalities within a structured rehabilitation setting. This pilot study was designed to address this issue.

Its primary objective was to evaluate the feasibility of an in-hospital LSVT BIG training program by assessing recruitment capability, compliance, and adherence to the program. The study’s secondary objectives were to evaluate the effects of the LSVT BIG training on gait, balance, and functional outcomes, as well as respiratory function after training in comparison with a progressive structured care (SC) rehabilitation program. The impact on patients’ overall quality of life was also evaluated.

2. Materials and Methods

2.1. Design of the Study

This was a feasibility study of an in-hospital high-intensity training program in patients with Parkinson’s Disease with exploratory assessments on functional and respiratory outcomes in comparison with a structured rehabilitation treatment (proof of concept). Exploratory assessments were collected in a prospective randomized manner, adhering to the ethical guidelines outlined in the Declaration of Helsinki.

All participating patients provided informed written consent to be part of the study. The study protocol received approval from both the Institutional Review Board and the Maugeri Ethical Committee (CE 2571, approved on 7 September 2021).

2.2. Patients

Patients attending the ICS Maugeri Institute of Castel Goffredo as outpatients (day hospital) for two years were randomized 1:1 in this pilot study to two groups of short-term training (four weeks). Figure 1 shows the participants’ flow and timeline through the study. Consolidated standards of reporting trials (CONSORT) flow chart was used to describe the flow.

In the current study, we included male and female patients diagnosed with mild-to-moderate stage Parkinson’s disease (Hoehn & Yahr 1–3) [21]. These patients consecutively attended the Institute of Castel Goffredo between September 2021 and August 2023.

We excluded patients with advanced PD (HY ≥ 3.5), cognitive impairment on Mini Mental State Examination [22] assessment (MMSE < 25), previous severe psychiatric disorders, atypical parkinsonism, stroke outcomes, orthopedic conditions, or those requiring long-term oxygen therapy (LTOT). At enrollment, eligible patients underwent a comprehensive clinical neurological evaluation to confirm their suitability for the study. Their existing drug therapy was also recorded and optimized at this time.

2.3. Rehabilitation Program

All patients with clinical indications for outpatient rehabilitation participated in a comprehensive program. The rehabilitation program included a structured exercise component (60 min per day, five days per week), occupational therapy (30 min per day, five days per week), or exercise for cognitive stimulation (30 min per day, five days per week), for a total of 450 min per week.

Four-week activities for both groups were categorized as general physiotherapy (gait, balance, and functional) or aerobic and were tailored to each group’s specific protocol. To ensure a comparable total amount of therapy received between groups, consistent with our Institute’s structured outpatient rehabilitation program (typically involving five days per week for four weeks), the physical therapist (PT) led a muscle relaxation activity for both groups on the fifth day of the second and third weeks. During these sessions, the PT assessed the participants’ understanding of and motivation for the activities.

2.4. Study Groups and Specific Training Programs

The structured exercise component, administered during the “ON” phase of their pharmacological effect (when their medication was most effective), was one of two options that randomly (1:1) assigned eligible patients to one of two groups: the structured care (SC) group or the LSVT BIG group (Figure 2). The randomization was performed using a block randomization criterion of 4, generated by https://www.randomizer.at/ (29 September 2025), and managed by a professional not involved in the study. The group assignment was not disclosed to the principal investigator at the time of the medical examination. It was only communicated to the PTs involved in LSVT training and standard care for competence after the clinical evaluation and test execution. Thus, the principal investigator was blind to allocation, while PTs involved in the training were not.

On the first day of the first week (baseline), the same PT (who was blinded) assessed all participants with the tests required by the study. Starting from day two and for the duration of the four-week program, patients in both groups received specific training in a 1:1 PT-to-patient ratio.

Upon discharge (the last day of the fourth week), the test assessments were followed up by different PTs (no blinding) according to the training session. As regards neuropsychological tests (subjective outcomes), these were carried out with the neuropsychologist in blindness.

2.4.1. Progressive Structured Care (SC) Rehabilitation Program

Patients assigned to the structured care (SC) group followed a standard rehabilitation program intentionally designed to match the exercise intensity of the LSVT BIG program. The SC training program began on day 2 of the first week and combined functional and aerobic elements.

The rehabilitation sessions consistently followed a specific sequence beginning with free-body functional activities, which included both passive mobilization and active-assisted movements (general physiotherapy), that were then followed by aerobic activity using treadmill training. Specifically, SC involved the following functional activities:

• Motor rehabilitation focusing on gradual, full-range passive mobilization exercises for the trunk and limbs, performed on a physical therapist’s couch.
• Active or assisted mobilization exercises, such as active head and trunk extension exercises, muscle activation for upper and lower limbs, and stretching exercises.

During the initial treatment phase, patients performed active and passive mobilization of the main joints in various positions (supine, prone, lateral decubitus, and quadrupedal), incorporating stretching exercises for the posterior and anterior kinetic chains.

• Trunk rotation exercises, sit-to-stand and stand-to-sit transitions, and balance exercises in both sitting and standing positions.

Postural control activities were performed in the standing position first, followed by straightening exercises, squats, lunges, and walking on the spot, with an increase in workload. Before the aerobic training, motor coordination exercises of increasing complexity were performed, as well as balance activities with and without visual feedback, within the limits of the patient’s initial clinical presentation.

• Aerobic training consisted of walking on a treadmill, with speeds ranging from 1.8 to 3.5 km/h. The intensity of the treadmill training was carefully monitored using the Borg Fatigue Scale [23] (range 5–7), using Maltais protocol [24], ensuring it was as intensive as the LSVT BIG program. The heart rate was constantly monitored to be less than 90% of predicted (220 beats per minute—age). This aerobic component also incorporated a progression in intensity throughout the program.

While the types and order of exercises remained consistent throughout the study, the time dedicated to each type of activity changed weekly, as follows:

• Time for free-body functional activity decreased from 45 to 30 min (5 min less each week, starting from week 1 to week 4).
• Time for aerobic treadmill activity increased from 15 to 30 min (5 min more each week, starting from week 1 to week 4).
• By week 4, an equal amount of time (30 min) was spent on both types of activities (Figure 2).

Moreover, once at home, the patient was free to do exercise.

2.4.2. LSVT BIG

LSVT BIG is a specialized training program that is especially effective for individuals in the early stages of Parkinson’s Disease (PD). The program focuses on making movements bigger and more intentional, which helps counteract the smaller, slower movements often seen in PD. LSVT BIG is structured around three key components:

• Functional Tasks and Structured Activities: Patients practice large, exaggerated movements with their whole body. This includes repetitive movements in various directions, like stepping and reaching, and gentle stretching. The goal is to maximize the amplitude of their movements.
• Goal-Directed Activities of Daily Living (ADLs): This component integrates those “bigger” movements into everyday tasks. Patients practice ADLs (like getting dressed or eating) with high-amplitude movements and an aerobic component to improve their overall physical endurance.

Both Functional Tasks and Structured Activities and Goal-Directed Activities of Daily Living can be considered in comparison to general physiotherapy (gait/balance/functional).

• The “BIG Walk” is a crucial aerobic part of the program. Patients are encouraged to walk with large, exaggerated movements of both their arms and legs. This walking exercise is personalized to each patient’s abilities and adapted to various real-world situations. For instance, they might practice changing direction, navigating narrow spaces, stepping over obstacles, or walking through doorways, all while maintaining those big, purposeful movements.

Moreover, although it is not the main focus as in LSVT LOUD, in the LSVT BIG program, the use of voice (loud, clear, and even shouted) is strongly recommended by the program during physical exertion to help the patient maintain range and intensity of movement (i.e., counting or accompanying exercises). It is intended as complementary and functional to the motor goal.

Patients participating in LSVT BIG received one-on-one, high-intensity training from a physical therapist (FE) who was specifically certified as an LSVT BIG instructor [16,17]. This personalized approach meant the PT trained together with the patient, providing immediate feedback on correct movements in real-time.

As described in the SC group, throughout the program, there was an adjustment in the time spent on different activities. Initially, more time was dedicated to functional activities, gradually decreasing from 45 to 30 min over the weeks (5 min each week). Conversely, the time spent on the “BIG Walk” (aerobic activity) increases from 15 to 30 min. By the fourth week, an equal amount of time was spent on both types of activities.

The intensity of the aerobic exercises was carefully monitored using the Borg fatigue scale [22], aiming for a rating of 5 to 7 to ensure patients were working hard enough without overexerting themselves (Figure 2). Heart rate was constantly monitored to be less than 90% of predicted (i.e., 220 beats per minutes—the patients’ age).

Moreover, once at home, the patient was free to apply the LSVT technique learned during the training in daily practice by performing the exercises using a DVD provided by the operator.

2.5. Measures

2.5.1. Primary Outcome

For this study, recruitment capability was defined as the number of patients who accepted the program with respect to the number of patients who fulfilled inclusion criteria; the compliance was evaluated as the number of patients who completed the four-week program and adherence with the program was strictly defined and measured by the number of patients who regularly attended all the rehabilitation sessions. Adherence was entirely supervised by a PT, and if a patient missed even one session, it resulted in classification as totally non-adherent to the study program.

Any refusal to participate, dropout, or absence from a session was recorded alongside the date of the event and the reason.

2.5.2. Secondary Outcomes

Measures were collected at baseline (T0) and at the end of in-hospital training (T1). Specifically, to maintain consistency, the same PT/medical doctor/neuropsychologist was responsible for the same tests and assessments throughout the study.

At T0, we collected anthropometric data (age, sex, weight, and body mass index), years from PD onset, and presence of main comorbidities (COPD, CHF, diabetes, hypertension, and others).

At T0 and T1:

(a)

Disease progression by motor experiences of daily living (UPDRSII, score from 0 to 52—lowest level of function) and PD-related motor examinations (UPDRSIII score from 0 to 132—lowest level of function) [25].

(b)

Gait speed calculated from the 10 min walking test [26] (gait speed = seconds to do the task/10 m distance covered) and freeze of gate [27] (FOG-Q, score from 0 to 108—lowest level of function).

(c)

Balance by Mini-BESTest [28] (Mini-BESTest, score from 0 to 28—highest level of function), Timed Up and Go test [29] as part of Mini-BESTest (Mini-BESTest-TUG, expressed in seconds to do the task), Timed Up and Go dual task as part of Mini-BESTest (Mini-BESTest-TUG cog, expressed in seconds to do the task with cognitive involvement) and Berg Balance Scale (score from 0 to 56—highest level of function) [30].

(d)

Lower limb strength by the 5 repetitions Sit-to-Stand test (5STS, expressed in seconds to do the task) [31].

(e)

Lower limb and upper limb exercise tolerance by 6 min walking test (6MWT) [32] and the Unsupported Upper Limb Exercise test (UULEX) [33], respectively. For 6MWT, the distance walked in meters was collected as % of predicted meters [34] calculated, and the Borg Fatigue scale [23] was recorded at the beginning and end of the test. UULEX maximum level reached and the total time of duration were recorded. For both tests, mean oximetry (SpO₂%) (Spirodoc, MIR Medical International Research, USA) was recorded.

(f)

Respiratory muscle strength by maximal inspiratory pressure (MIP) and maximal expiratory pressure (MEP) [35] (MicroRPM, Care-Fusion, Hoechberg, Germany), both expressed in cmH₂O and % of predicted value and respiratory function by peak cough expiratory flow was recorded (PCEF, expressed in L/min) [36] (Spirodoc, MIR Medical International Research, Suite O. Waukesha, WI, USA with Ambu Ultraseal mask, Ballierup, Denmark).

(g)

Quality of life by Parkinson Fatigue scale [37] with polytomous correction method (PFS, score from 0 to 5—lowest level of function) and Parkinson’s Disease Quality of Life Questionnaire-39 scale [38] with its Summary Index (PDQ39-SI, score% from 0 to 100%—lowest level of function).

We considered the Minimal Clinical Importance Difference available in PD or in other diseases/general population for UPDRSIII (−2.5 points) [39], MIP (+17.2 cmH₂0) [40], 6MWT (+30.5 m) [41], Berg Balance Scale (different delta points depending on baseline value) [39], Mini-BESTest (+4 points) [42], and Gait Speed (+0.16 m/s) [39].

2.6. Statistical Analysis

A per-protocol analysis was performed on patients who completed the rehabilitation program and attended the visit at the end of the rehabilitation (T1). Data were summarized as mean ± standard deviation (SD) or median and 1st–3rd quartile for quantitative variables and frequencies for categorical variables before undergoing statistical analysis with Graph Pad Software (Prism 4, San Diego, CA, USA) and R programming language (Vienna, Austria, 2015) [43].

To compare baseline measures and delta changes between the two groups, we used Wilcoxon’s non-parametric test and Pearson’s Chi-squared test (applying the Monte Carlo correction in the case of low numbers).

Calculation of the effect size for comparisons between means was performed by Cohen’s and between medians by rank biserial correlation. Effect size was retained small (d = 0.1), medium (d = 0.3), and large (d ≥ 0.7) [44].

The risk of association of the improvement over the MCID and the LSVT BIG group was evaluated by the odds ratio (OR) with Fisher’s test, considering all variables with an MCID from the literature, except for the Berg Balance. Median value of Berg Balance in the studied patients at baseline (52 points) was high (good balance state), so in the majority of cases, too close to the maximum value (56 points) to permit a change of >4 points. The ODDS ratio (improvement over the MCID, yes/no versus LSVT BIG/SC) was evaluated by the Fisher test.

The analysis of variance for repeated measures (ANOVA, group ∗ time interaction) was performed on all variables. If ANOVA showed statistical significance, the Holm–Bonferroni post hoc correction was used to evaluate differences between groups and time.

Correlations between variables were analyzed by Spearman’s correlation and were considered small, medium, and large effect sizes for values of 0.3, 0.5, and 0.6, respectively [44]. For all analyses, a value of p < 0.05 was considered statistically significant.

3. Results

A flow chart of the study is presented in Figure 1. Out of 48 patients with Parkinson’s disease assessed for eligibility, six did not fulfil the criteria due to other comorbidities, and two refused to participate for personal reasons. Forty participants were included and randomized to LSVT BIG group (n = 20) and SC group (n = 20).

3.1. Primary Outcome: Feasibility

The recruitment capability for the study was 40/42 subjects (95%), with only two patients who refused to participate. During the in-hospital program, another two participants dropped out (one from each group) due to personal reasons not dependent on the LSVT BIG or SC program. No adverse events, side effects, or discomfort were registered. Thus, compliance with the LSVT BIG program was 95% and measures were collected in 38 patients. All LSVT BIG patients were able to attend the hospital and complete the one-hour sessions five days a week as scheduled, without difficulty or interruption due to excessive fatigue. Adherence to the sessions was reported as 100%.

3.2. Sample Characteristics

Table 1. A comparison of the two training programs based on demographic and clinical data and functional measures of the entire population at baseline (T0).

	All (n = 38)	LSVT BIG (n = 19)	SC (n = 19)	p
Age, years	72 (68–76)	71 (70–76)	73 (67–76)	0.2125
Male, n (%)	25 (66%)	11 (58%)	14 (74%)	0.4950
BMI, Kg/m2	26.7 (4.4)	26.6 (5.0)	26.8 (3.8)	0.6827
Disease duration PD, years	5 (3–8)	7 (4–9)	4 (2–8)	0.2125
Comorbidities
COPD yes, n (%)	8 (8)	2 (11)	1 (5)	1.0000
CHF yes, n (%)	7 (18)	4 (21)	3 (16)	1.0000
Diabetes yes, n (%)	7 (18)	4 (21)	3 (16)	1.0000
Hypertension yes, n (%)	20 (53)	11 (58)	9 (47)	0.7459
Others yes, n (%)	10 (26)	4 (21)	6 (32)	0.7141
Levodopa, yes n (%)	36 (95)	19 (100)	17 (89)	0.4865
Dopamine agonist, yes n (%)	23 (61)	13 (68)	10 (53)	0.5076
IMAO B, yes n (%)	18 (47)	9 (47)	9 (47)	1.0000
MMSE, score	28.4 (26.0–30.0)	28.3 (26.2–29.9)	29.0 (25.9–30.0)	0.6749
HY, score	2.0 (2.0–2.5)	2.0 (2.0–2.5)	2.0 (2.0–2.5)	0.8586
UPDRSII, score	6 (3–9)	7 (4–9)	5 (3–9)	0.4112
UPDRSIII, score	11 (8–16)	10 (8–16)	11 (9–17)	0.9767
Gait speed #, m/s	1.3 (1.2–1.6)	1.3 (1.2–1.7)	1.3 (1.2–1.6)	0.9767
FOG-Q, score	3.5 (0.3- 9.8)	4.0 (1.5–10.5)	2.0 (0.0–9.0)	0.4169
MIP, cm H2O	47 (18)	43 (15)	51 (21)	0.2119
MIP, % pred	52 (19)	50 (16)	54 (21)	0.4831
MEP, cm H2O	68 (22)	68 (22)	67 (22)	0.8829
MEP, % pred	40 (12)	42 (13)	38 (11)	0.3282
PCEF, L/min	274 (130)	246 (104)	301 (150)	0.1955
Berg Balance, score	52.0 (49.0–54.0)	50.0 (48.0–53.5)	53.0 (50.0–54.0)	0.5184
MB, score	20.3 (4.0)	20.2 (3.7)	20.3 (4.5)	0.9373
MB-TUG, s	9.3 (7.6–10.4)	9.2 (7.4–10.4)	8.3 (7.8–10.5)	0.3606
MB-TUG cog, s	11.7 (3.9)	11.5 (3.6)	11.9 (4.2)	0.7604
5 STS, s	14.2 (3.8)	13.9 (4.0)	14.4 (3.7)	0.6842
6MWT, meters	367 (103)	370 (106)	365 (104)	0.8776
6MWT, % pred	82 (68–92)	83 (69–90)	79 (66–92)	0.8491
6MWT, SatO2 mean, %	97 (97–98)	97 (96–98)	97 (96–98)	0.7901
6MWT, Exercise DES, n (%)	2 (5.3)	1 (5.3)	1 (5.3)	1.0000
6MWT, Borg Fatigue pre, score	1 (0–1)	0 (0–1)	1 (0–1.5)	0.8289
6MWT, Borg Fatigue post, score	4.5 (3–6)	4 (2.5–6)	5 (3–7)	0.2305
UULEX time, s	278 (120)	259 (118)	297 (121)	0.3295
UULEX level, score	5 (2)	5 (2)	5 (2)	0.3843
UULEX, SatO2 mean, %	97 (96–98)	97 (96–98)	97 (97–98)	0.7010
UULEX Exercise DES, n (%)	0 (0.0)	0 (0.0)	0 (0.0)	1.0000
PFS, score	2.4 (0.8)	2.4 (0.7)	2.4 (0.8)	0.8949
PDQ39-SI, score%	19.5 (12.3– 26.8)	19.8 (17.0–24.7)	17.9 (10.3–31.4)	0.5395

Legend: values expressed as mean (SD) or median (1–3° quartile) or number (n) and percentage (%). CI 95% = confidence interval at 95%; LSVT = Lee Silverman Voice Treatment; SC = structured care; BMI = body mass index; MMSE = Mini-Mental State Examination; HY = Hoehn and Yahr staging scale; UPDRS = Unified Parkinson’s Disease Rating Scale; FOG_Q = Freezing of Gait Questionnaire; 6MWT = 6 min walking test; % pred = percentage of predicted value; SatO₂ = oxygen saturation; MIP = maximal inspiratory pressure; MEP = maximal expiratory pressure; PCEF = peak cough expiratory flow; MB = Mini-BESTest; TUG = Timed Up and Go test; TUG cog = Timed Up and Go dual task (cognitive involvement); 5 STS = 5 repetitions Sit-to-Stand; UULEX = Unsupported Upper Limb Exercise test; PFS = Parkinson Fatigue scale (polytomous method score); PDQ39_SI = Parkinson’s Disease Quality of Life Questionnaire-39 scale_ Summary Index; # gait speed calculated from 10 m walking test. Per-protocol analysis was performed on 38 patients who completed the program.

describes the PD anthropometric and clinical characteristics of the recruited sample, with a median age of 72 years, 66% male, time since PD diagnosis ranging from 1 to 24 years (median 5 years), and Hoehn and Yahr scores ranging from 1 to 3 (median 2). Demographic variables, disease duration, PD stage, and comorbidities were similar between groups (p > 0.05 for all) (Table 1). Although PD patients were in the early stages of the disease with gait/balance/functional scores more similar to those of elderly subjects in good physical condition, they presented a pronounced inspiratory and expiratory muscle weakness, weak cough, and a mild reduction in their ability to tolerate physical effort during ambulation and upper extremity tasks.

3.3. Secondary Outcomes: Pre-to-Post Changes Between Outcomes

The trend of the measures between T0 and T1 was analyzed with the ANOVA for repeated measures and is presented graphically in Figure 3 for all evaluations with at least one significant p-value (for group, time, or group ∗ time). PDQ39 and PFS were not represented due to a total absence of significance. ‘Time’ was independently significant for all measures except 5-STS. ‘Group’ was never independently significant. The interaction ‘group ∗ time’ was significant for the Mini-BESTest score, UULEX, and gait speed (panels i, l, and c). Post hoc tests revealed that T1 differed significantly from T0 for both LSVT and SC in Mini-BESTest score (panel i) and 6MWT (panel k), while for LSVT only in gait speed, FOG-Q, MIP, MEP, PCEF (panels d, e, f), Berg Balance (panel h), and UULEX (panel l). All between-group comparisons at the same times (T0 and T1) were not significant.

Table 2. A comparison of the delta changes for the T1−T0 assessments in the two groups.

Variables	LSVT BIG (n = 19)	SC (n = 19)	p	Effect Size (CI95%)
UPDRSII, score	−2.0 (−3.0; −0.5)	−1.0 (−3.5; 0.0)	0.7240	−0.059 (−0.372; 0.266)
UPDRSIII, score	−2.9 (5.2)	−2.9 (4.4)	1.000	0 (−0.636; 0.636)
Gait speed #, m/s	0.15 (0.03; 0.29)	0.02 (−0.04; 0.12)	0.0295 *	0.355 (0.040; 0.606)
FOG-Q, score	−1.0 (−3.0; 0.0)	0.0 (−3.0; 0.0)	0.3571	−0.149 (−0.448; 0.179)
MIP, cm H2O	11 (5; 17)	2 (−2; 11)	0.0877 °	0.358 (0.043; 0.608)
MEP, cm H2O	9 (3; 15)	2 (−3; 13)	0.1561	0.232 (−0.095; 0.514)
PCEF, L/min	48 (17; 79)	17 (−7; 43)	0.0343 *	0.346 (0.029; 0.599)
Berg Balance, score	0.0 (−4.2; 1.0)	−1.3 (−5.1; 5.4)	0.0933 °	0.270 (−0.054; 0.543)
MB, score	3.9 (2.4)	1.8 (2.7)	0.0129 *	0.822 (0.153; 1.480)
MB-TUG, s	−1.5 (−2.1; −0.6)	−0.9 (−1.3: −0.2)	0.1217	−0.253 (−0.530; 0.072)
MB-TUG cog, s	−1.6 (−3.7 −0.4)	−0.9 (−1.8; 0.0)	0.1025	−0.268 (−0.541; 0.057)
5 STS, s	−2.0 (−3.1; −0.9)	−0.8 (−2.4; 0.9)	0.0876 °	−0.279 (−0.550; 0.044)
6MWT, meters	60 (35; 83)	5 (−3; 54)	0.0113 *	0.412 (0.106; 0.647)
6MWT, SatO2 mean, %	0.0 (0.0; 0.5)	0.0 (−1.0; 0.5)	0.1428	0.212 (−0.112; 0.501)
UULEX level, score	1.0 (1.0; 2.5)	0.0 (0.0; 1.0)	0.0002 **	0.561 (0.294; 0.747)
UULEX time, s	71 (52; 163)	9 (−2; 47)	0.0004 **	0.573 (0.310; 0.755)
UULEX, SatO2 mean, %	1.0 (0.8)	0.3 (1.2)	0.0481 *	0.686 (0.027; 1.337)
PFS, score	0.0 (0.7)	−0.3 (0.6)	0.0836 °	0.460 (−0.188; 1.102)
PDQ39 SI, score%	0.0 (−4.2; 1.0)	−1.3 (−5.1; 5.4)	0.8839	−0.026 (−0.296; 0.343)

Legend: values expressed as mean (SD) or median (1° 3° quartile). LSVT = Lee Silverman Voice Treatment; SC = structured care; UPDRS = Unified Parkinson’s Disease Rating Scale; FOG-Q = Freezing of Gait Questionnaire; SatO₂ = Oxygen saturation; PCEF = peak cough expiratory flow; MIP = maximal inspiratory pressure; MEP = maximal expiratory pressure; MB = Mini-BESTest; TUG = Timed Up and Go test; TUG cog = Timed Up and Go dual task (cognitive involvement); 5 STS = 5 repetitions Sit-to-Stand; 6MWT = 6 min walking test; UULEX = Unsupported Upper Limb Exercise test; PFS = Parkinson Fatigue scale (polytomous method score); PDQ39_SI = Parkinson’s Disease Quality of Life Questionnaire-39 scale_ Summary Index; # gait speed calculated from 10 m walking test. ** p < 0.01; * p < 0.05; ° 0.05 < p < 0.10. Effect size was calculated by Cohen’s d for the mean and the rank biserial correlation for median values.

compares the changes between the groups from pre to post (delta T1−T0). Several parameters (in bold) showed a difference between groups: the UULEX (level and time) showed strong statistically significant differences (p < 0.001); UULEX SatO₂ mean, gait speed, PCEF, Mini-BESTest and 6MWT differences showed statistical significance with p < 0.05, while MIP, Berg Balance, 5 STS and PFS showed differences close to significance. For all of the aforementioned measures, the LSVT BIG delta change was consistently greater than the SC delta. It should be noted that the mean delta T1−T0 was equal to or greater than the MCID for UPDRS III (for both LSVT BIG and SC, with a delta UPDRS III of −2.9 points), for gait speed (for LSVT BIG only, with a delta gait speed of 0.18 m/s versus 0.07 m/s for SC), for 6MWD (for LSVT BIG, with a delta 6MWD of 62.9 m, and for SC, with a delta 6MWD of 34.6 m), and for Mini-BESTest (for LSVT BIG only, with a delta Mini-BESTest of 3.9 points, very close to 4 points versus 1.8 for SC).

For the remaining measures, the LSVT BIG group showed comparable improvement (delta T1−T0) to the SC group, with no significant differences between groups.

The estimation of the effect size evaluated on data from 38 PD patients demonstrated that only UULEX (level, time, and SatO₂) and Mini-BESTest exhibited a medium and large clinical effect, respectively.

Table 3. Evaluation of the probability that the group LSVT BIG will have a gain (delta change T1−T0) in some variables that is greater than the corresponding MCID.

Variables	LSVT BIG (n = 19)	SC (n = 19)	OR	95% CI	p
Delta T1-T0 UPDRS III ≤−2.5 points, n (%)	9 (47%)	9 (47%)	1.00	0.23; 4.28	0.3269
Delta T1-T0 Gait speed # ≥0.16 m/s, n (%)	9 (50%)	3 (16%)	4.60	0.87; 32.89	0.0789 °
Delta T1-T0 MIP ≥17.2 cmH2O, n (%)	5 (26%)	4 (21%)	1.33	0.23; 8.18	1.0000
Delta T1-T0 MB ≥4 points, n (%)	12 (63%)	3 (16%)	8.54	1.63; 62.54	0.0069 **
Delta T1-T0 6MWD ≥30.5 m, n (%)	14 (74%)	8 (42%)	3.71	0.82; 19.17	0.0991 °

Legend: number (n) and percentage (%), CI 95% = confidence interval at 95%; LSVT = Lee Silverman Voice Treatment; SC = structured care; UPDRS = Unified Parkinson’s Disease Rating Scale; MIP = maximal inspiratory pressure; MB = Mini-BESTest; 6MWD = 6 min walking distance; # gait speed calculated from 10 m walking test. ** p < 0.01; ° 0.05 < p < 0.10.

shows the number of patients in the two groups with a T1−T0 change only in variables that achieved or exceeded the relevant known MCID. The only notable odds ratio was observed for Mini-BESTest: patients in the LSVT group were 8.54 times more likely to exceed the MCID for Mini-BESTest than those in the SC group.

3.4. Correlations

Spearman correlations between UPDRS II (ADL component) and UPDRS III (motor component) at T1 and measurements at T0 are shown in Table S1 (Supplementary Materials). In general, UPDRS II and UPDRS III correlated with the same parameters at T0, UPDRS III having stronger correlations than UPDRS II in most cases. Moderate correlations were observed between UPDRS II and UPDRS III with MIP % of predicted, Berg Balance, MB-TUG, UULEX level and time, and 6MWT % of predicted. Only UPDRS III, and not II, presented a moderate correlation with FOG-Q.

4. Discussion

The goal of this study was to determine if in-hospital LSVT BIG training was a feasible and well-followed intervention for patients with Parkinson’s disease, especially when compared to a progressive structured care rehabilitation program with similar frequency and intensity. The study’s key findings demonstrated the feasibility of in-hospital LSVT BIG training, evidenced by a good capability recruitment, a strong safety profile, and full adherence. Accordingly, motor functional evaluations from these patients could be informative in terms of patients’ improved mobility, gait and balance, and effort tolerance, as well as, for the first time, respiratory function evaluations, after the in-hospital training in both groups (Table 2).

Significant improvements in mobility and balance, as measured by the Gait and Mini-BESTest, were exclusively observed in the LSVT BIG training group. This group also uniquely demonstrated improved lower and upper limb effort tolerance, as indicated by 6MWT, UULEX time, and UULEX SatO₂. Additionally, PCEF showed a significant difference, favoring the LSVT BIG group. Quality of life was unchanged across both groups. However, these differences may not be due to a methodological LSVT BIG superiority but could be influenced by other factors, such as higher patient motivation, ability to repeat exercises at home, or individualized therapist attention.

4.1. Primary Outcome

Drawing upon our experience, we posit that the LSVT BIG intervention is feasible in the context of routine clinical practice, despite the constraints imposed by limited resources (i.e., a single licensed physical therapist) and by the fact that it is carried out in a small territorial hospital located in the Po Valley. Remarkably, we were able to successfully deliver the study and rehabilitation intervention, with 95% consent to participate and 95% to conclude the whole program in total safety.

This result can be regarded as a success from different perspectives. Firstly, it is evident from the extant literature that LSVT BIG training has been predominantly used in the home environment, in non-purely rehabilitative or hospital-based contexts. This is mainly because the main goal of maintaining the frequency of the standard protocol (4 consecutive days per week for 4 weeks) is challenging for most patients, particularly in terms of access to dedicated rehabilitation clinics for LSVT BIG treatment; thus, a different approach by telemedicine has been developed [20,45]. In our particular case, the geographical location of the rehabilitation institute and the commitment of the patients have been instrumental in overcoming this challenge and ensuring the highest possible level of patient attendance.

Secondly, unlike what often happens in community-based models of PD rehabilitation [8], where the benefits of the intervention or its maintenance over time are limited by reducing the intensity of rehabilitation programs, fragmentation, and inadequate scheduling, in the current study, adherence of patients in a mild–moderate stage of disease to the in-hospital programs was 100%. This result allows us to proceed to give information in a discrete population of patients about the comparison between the two groups of study.

4.2. Secondary Outcomes

The neuromotor data collected from the study participants at enrolment are consistent with results from earlier research by our group [46]. Concerning the UPDRS III scores, a measure of motor symptoms in Parkinson’s Disease, recorded in our study, these were lower than what is typically reported for Parkinson’s Disease patients in other studies [19,20,47,48]. Due to these lower scores, our basal findings were not directly comparable to those from others. This discrepancy further emphasizes that the pathology in our study population was at its onset compared to what was observed in the populations studied by other authors. However, the UPDRS III score improved similarly by the end of the rehabilitation period, in line with the trend observed in Shable’s research [19]. The effects of LSVT BIG therapy on motor function have been shown to be superior to those of Nordic walking or home-based training [17]. Nevertheless, the current study, which compares with a structured rehabilitation in-hospital program, observed no statistical difference between groups, in line with Shable’s findings [19].

The current study provides compelling evidence of rehabilitative improvement in further parameters such as gait and balance and effort tolerance in both participant groups following their respective training programs. Improvements were distinctly noted in gait speed and scores on the Mini-BESTest, indicating enhanced dynamic stability and walking performance. The findings are in line with Mehdizadeh et al. [49], that suggests the Mini-BESTest is a responsive outcome measure for PD.

Our findings on endurance uniquely demonstrate, for the first time, an increase in the effort tolerance of both upper and lower limbs. This was quantitatively measured by improvements in the 6MWT, reflecting enhanced lower-limb endurance, and the UULEX test, indicating greater upper body work capacity. All these observed improvements, while varying in magnitude, strongly support the hypothesis that exercise-dependent plasticity serves as the primary mechanism underpinning the therapeutic effects of physiotherapy. This suggests that the brain’s remarkable capacity for adaptation and reorganization through physical activity is fundamental to achieving rehabilitative gains, irrespective of the specific exercise modality employed. This concept aligns with the existing literature [8].

Our specific gait/balance findings are highly consistent with results reported in other LSVT BIG studies [20,47,48], thereby reinforcing the broader scientific consensus on the benefits of rehabilitation in this population.

As for lower limb effort tolerance using the 6MWT and the Upper Limb Exercise using the UULEX test, the present study provided valuable data. Direct comparisons with the existing literature are difficult because no publication in the literature combines 6MWT and UULEX in patients with Parkinson’s disease during the LSVT BIG program, mainly focusing on balance and motor function. Indeed, only one study evaluated the 6MWT test [50], providing evidence for an improvement at the end of the program; on the contrary, the UULEX test is mainly used in cardio-respiratory medicine and does not appear in studies on LSVT BIG.

Despite this, a significant finding emerged because the LSVT BIG group showed, in 6MWT, a significantly greater improvement over the Minimal Clinically Important Difference (MCID) (mean 62.9 m) with respect to the standard care (SC) control group (mean 34.6 m). Such a comparison for upper limbs using UULEX was not possible due to the unavailability of an MCID for this test. Indeed, UULEX improvement in LSVT BIG (median 71 s) was significantly greater than UULEX improvement in SC (median 9 s). This strongly suggests that LSVT BIG leads to superior functional performance gains.

One possible explanation is that although we did not monitor physical activity at home during the hospital rehabilitation program, the results might depend on the fact that motivated patients were free to repeat the LSVT BIG exercises at home using a DVD, effectively doubling their training/day and consequently achieving better results. Moreover, when directly compared to the SC group, the LSVT BIG group demonstrated statistically significant advantages not only in these effort tolerance parameters but also in peak cough expiratory flow (PCEF), a key respiratory measure.

In the field of respiratory diseases, the link between upper limb training and respiratory muscle function is deeply studied because arm training can contribute to improvement in respiratory muscle performance, and the clinical benefit of arm training is most evident in improved exercise tolerance and reduced dyspnea during activities involving the upper limbs in respiratory patients [51,52].

Physiologically, unsupported arm activity increases ventilatory demand and recruits both respiratory and accessory muscles, including those of the shoulder girdle, which may contribute to improved ventilatory muscle performance over time [53].

In particular for PCEF, expiratory muscle weakness impairs cough, increasing the risk of atelectasis. In our LSVT BIG training, the upper limb exercises, in addition to abdominal muscles exercises, could have had a role in the expiratory muscle strength improvement and cough effectiveness, whereas arm training does not directly address these mechanisms [54,55]. Nevertheless, future studies will need to confirm this effect and compare it with traditional respiratory muscle training [56,57,58,59].

Concerning QOL (PDQ-39 scores), we did not observe direct or indirect evidence of LSVTBIG^® training, in agreement with the Cochrane findings [7].

4.3. Clinical Implication

The main issue is the lack of comparative studies with other intensive therapy forms. These novel findings underscore the crucial role of frequency, intensity, time, and type of exercises in a rehabilitation program. They also prompt us to consider if certain physiotherapy interventions, like LSVT BIG, may offer distinct advantages for specific outcomes, as we observed for the dynamic balance (measured by Mini-BESTest) or effort tolerance compared to a structured care program, balanced for time, frequency, and intensity.

Additionally, for the purpose of generalizing the results and personalizing the training, it is essential to note that the intensity of the ‘endurance’ component was defined using the Borg scale [23] with a target range of Borg 5 to 7. Although this is a widely used method for adjusting load progression—also accepted in certain guidelines [24]—our chosen intensity level should be carefully evaluated for each patient. This is because even mild cognitive impairments may affect the correct use of the scale. In some cases, lower intensity levels may be required (e.g., in the presence of cardiac comorbidities). Close monitoring of the physiological response is particularly recommended during the initial sessions.

The benefits seen with LSVT might stem from its specific exercise methodology, which heavily engages large movements of the trunk and upper limbs, integrating the use of voice to reinforce motor commands. However, this combined training style, together with the patient’s motivation and ability to repeat exercises at home, or the therapist personalized attention, could have led to a more pronounced improvement in LSVT BIG when compared to traditional training methods. Indeed, motivation has been demonstrated to promote better adherence and greater symptomatic benefits in PD [60]. Therefore, these aspects require confirmation.

4.4. Limitations

The study’s main weaknesses are related to the high risk of bias (the lack of PTs’ blinding, and allocation concealment); potential confounding (participants were given a DVD for uncontrolled home practice); the potential for therapist allegiance bias; and the risk of false positive findings.

While LSVT BIG training shows promise, a key limitation to its widespread application appears to be the stage of PD, mainly during the initial mild–moderate phases. The literature has not explored the intensive nature of the program in individuals with more advanced stages (Hoehn and Yahr [HY] ≥ 3.5), and they might not achieve the same improvements as individuals in early-moderate stages who participated in our research. This observation is consistent with other studies [17,19,20,47], including one [20] where LSVT BIG was compared to telerehabilitation and found to be preferred for improving dynamic balance and activity status specifically in early-stage PD.

As this was a pilot study, a formal sample size calculation was not performed. Consequently, the statistical significances reported for time–group comparisons are not broadly generalizable to the wider PD population. The relatively small sample size (possibility of Type II error) for patients from a single site, coupled with the inherent high variability of the evaluated variables, likely prevented us from detecting statistical significance in some instances, even where visual inspection of median comparisons suggested potential differences between time points or groups.

The current paper is limited to considering the pre-to-post effect on rehabilitation. Further studies will assess whether the effects observed during hospitalization persist after patients return home.

4.5. Strengths

This feasibility study, like others reported in the literature, is unique because of (a) a randomization into two groups with balanced SC treatment on LSVT BIG, (b) a discrete number of patients analyzed, (c) a large number of motor functional measures collected, and (d) effort tolerance and respiratory evaluations as novelty.

5. Conclusions

This study shows that in-hospital LSVT BIG training for people with Parkinson’s disease is feasible, demonstrating good recruitment rates, a strong safety profile, and perfect adherence. Compared to a structured intensive care rehabilitation program of a similar frequency and intensity but a different type of exercise, LSVT BIG training seems to give better performance in dynamic equilibrium and effort tolerance. Due to the exploratory nature of the study on the secondary outcomes, these results are preliminary and need to confirm the role of patients’ motivation and exercise at home and require validation over time. Further RCT studies comparing LSVT BIG with protocols of similar intensity are also needed.