Low-intensity vibration to reduce symptoms and improve physical functioning in cancer survivors with chemotherapy-induced peripheral neuropathy: A pilot randomized trial

Abstract

Purpose

Chemotherapy-induced peripheral neuropathy (CIPN) can have deleterious effects on mobility and quality of life in people with cancer. Vibration therapy shows promise as a CIPN intervention but is understudied. We investigated the feasibility and preliminary efficacy of low-intensity vibration (LIV) in cancer survivors with CIPN.

Patients and Methods

We conducted a pilot randomized controlled trial in adult cancer survivors with persistent CIPN symptoms. Participants were randomized to twice daily LIV sessions (10 minutes/session; 30Hz, 0.4g) for 12 weeks or usual care (UC). We assessed feasibility by accrual, retention, adherence, and adverse event reporting. We evaluated preliminary efficacy by changes in patient-reported CIPN symptoms (FACT/GOG-Ntx), pain (Brief Pain Inventory), fatigue (PROMIS Fatigue) and physical functioning (LLFDI) and objectively measured physical functioning (chair stand time, gait speed), stability (postural sway), and mobility (Timed up-and-go). Linear regression models were used to generate effect size estimates (Cohen’s d).

Results

We accrued 95% of our target sample (n=38, mean age: 62.6 ± 9.9 years, 89% female, median time since chemotherapy completion: 18 (6–39) months), with 20 participants randomized to LIV and 18 to UC. Trial retention was 97% and mean adherence to LIV was 77% ± 18%. There were no serious adverse events. Compared to UC, LIV participants reported greater improvements in sensory neuropathy symptoms (LIV: +1.4 ± 3.3 points, UC: +0.2 ± 2.8 points; Cohen’s d=.45) and basic lower extremity function (LIV: +5.3 ± 8.5 points, UC: −0.7 ± 9.2 points; Cohen’s d=.80), with moderate-to-large effect sizes for changes in stability, mobility, and gait (Cohen’s d=.60–.66).

Conclusion

LIV is safe, feasible, and shows preliminary efficacy for CIPN symptom relief and improving physical functioning in cancer survivors with CIPN.

BACKGROUND

Chemotherapy-induced peripheral neuropathy (CIPN) is a common and debilitating side effect of neurotoxic chemotherapy.¹ CIPN is characterized by sensory and/or motor symptoms in the hands and/or feet, often leading to gait and balance impairments that increase disability and fall risk.^2,3 CIPN occurs in up to 90% of adults with cancer during neurotoxic chemotherapy and may persist in upwards of 50% of cancer survivors long-term.^4–6 Yet, prevention and treatment options for CIPN are limited. The American Society of Clinical Oncology guidelines do not endorse any evidence-based approaches to prevent CIPN except for neurotoxic chemotherapy dosing modifications in patients who develop intolerable symptoms and/or functional impairment.^7–9 Intermediate-grade evidence supports using duloxetine to treat CIPN-related symptoms; however, medications come with side effects and costs. Thus, non-pharmacologic approaches to manage CIPN and its harmful sequelae are needed.

Vibration therapy is a non-pharmacologic treatment modality that passively imparts oscillating mechanical signals to an individual standing on a vibrating platform.¹⁰ Vibration stimulates neuromuscular activity and proprioceptive pathways and may inhibit pain pathways^11,12. In adults with diabetic peripheral neuropathy, vibration therapy improves muscle strength, balance, touch sensation, and neuropathic pain.^13–16 Early phase trials suggest that vibration may improve pain and physical functioning in cancer survivors with CIPN. Current vibration modalities deliver accelerations above 5g (where 1g is Earth’s gravitational field), exceeding the International Organization for Standardization (ISO) threshold limits (ISO-2631)¹⁷ for safe durations of vibration exposure for healthy adults.^18–20, restricting its clinical and practical utility.^17,21 Thus, alternative vibration modalities must be considered.

Low-intensity vibration (LIV) is a safer alternative to current vibration platforms that may have similar clinical benefits, including improvements in balance, strength, and mobility in older adults.^22,23 LIV platforms deliver high-frequency, low-amplitude (<1g) mechanical stimulation at intensities deemed safe for several hours of daily use per ISO guidelines.¹⁷ LIV has not been studied as a treatment for CIPN. Thus, we conducted a pilot controlled trial in adult cancer survivors with persistent CIPN symptoms.²⁴ Our primary aim was to determine LIV’s feasibility (accrual, retention, adherence) and safety. Our secondary aims assessed LIV’s preliminary efficacy on CIPN-related clinical outcomes, including patient-reported outcomes (PROs) and objective assessments of physical functioning, stability, and mobility.

METHODS

Study design and setting

We conducted a 2-arm, parallel 1:1 assignment pilot RCT comparing 12 weeks of LIV to usual care (UC) at Oregon Health & Science University (OHSU). The OHSU Institutional Review Board approved the trial (#18342) which is registered on Clinicaltrials.gov (NCT04170075). Written informed consent was obtained from participants prior to data collection.

Study sample and recruitment

Participants were cancer survivors with persistent CIPN symptoms. Recruitment occurred from 7/10/2019 to 9/7/2022, with interruption from 2020–2021 due to COVID-19. Participants were recruited by clinician referrals and the OHSU cancer registry. Staff screened for the following eligibility criteria: (1) stage I-III cancer; (2) chemotherapy completion ≥6 weeks and ≤5 years of enrollment; (3) self-reported numbness and/or tingling in the feet; (4) age ≥18 years; (5) weight ≤275 pounds (i.e., platform weight limit); (6) inability to stand for 10 minutes unassisted and/or conditions that confound measures of change in stability and mobility. We aimed to enroll a sample size ≥ n=30 to assess feasibility outcomes and provide preliminary data to determine effect size estimates for future trials.

Randomization

Participants were randomized 1:1 to LIV or UC using a computer-generated (MS Excel) random numbers table created by the study biostatistician. Randomization was stratified by time since chemotherapy completion (≤1 year vs. >1 year) with a block size of 4. Group assignments were placed into sealed envelopes before enrollment and opened by participants after baseline testing. Study staff were unblinded to the treatment group after the baseline assessment.

LIV Intervention

Participants randomized to LIV were prescribed two home-based 10-minute LIV sessions daily for 12 weeks. Participants stood barefoot on a vibration platform (Marodyne LiV^®, BTT Health GmbH, Greifenberg, Germany) that applies barely perceptible high-frequency, low-amplitude vibrations (30 Hz, 120-micron vertical displacement, 0.4g peak-to-peak acceleration).

UC

UC participants were encouraged not to change their physical activity habits during the study period.

Procedures

Baseline demographics and history were collected by self-report. Patient-reported and objective measures were collected at baseline and after 12 weeks. Surveys were administered electronically (REDCap). Participants received $20/visit to offset time and travel.

Primary outcomes

Feasibility and safety

We defined feasibility as 100% trial accrual within a 4-month recruitment period, ≥80% trial retention, and ≥75% adherence to LIV sessions over the 12-week study period. We assessed accrual from enrollment. We defined trial retention as the percentage of enrolled participants who provided any outcomes data at baseline and 12 weeks. Intervention retention was defined as the percentage of persons who completed >50% of prescribed LIV sessions and did not withdraw from the study. Adherence to prescribed LIV sessions was assessed by participant diaries.

We defined safety as the absence of SAEs attributable to LIV using patient-reported adverse events (AEs) collected by weekly email surveys and patient reports via phone, email, and/or at the final testing visit. A blinded member of the study team (KWS) reviewed all AEs and graded them as mild (did not require medical treatment), moderate (resolved with medical treatment), or serious (life-threatening or disabling and requiring medical attention).

Secondary Outcomes

Patient-reported symptoms

CIPN

CIPN symptoms were assessed using the FACT/GOG-Ntx, an 11-item measure of the presence, severity, and interference of neuropathy symptoms.^25,26 The 4-item subscale assesses the presence and frequency of CIPN sensory symptoms. The trial outcome index (TOI; range: 0–100) adds questions about overall physical and functional well-being to the 11-item scale. Higher scores indicate less symptoms and better QOL. The minimal clinically important difference (MCID) is 0.3–0.4 points/item, or 1.2–1.6 points and 3.3–4.4 points for the 4-item and 11-item scales, respectively.²⁷

Pain

Global pain severity and interference were measured by the Brief Pain Inventory short form (BPI-sf). Higher scores indicate worse pain and greater interference.²⁸ Reduced pain severity of 10–20%, ≥30%, and ≥50% represent minimal, moderate, and substantial pain improvement, respectively.²⁹

Fatigue

Fatigue was measured by the PROMIS Fatigue - Short Form 8a, where higher scores indicate more fatigue.^30–33 The MCID is 3–5.^34,35

Physical functioning

Late-Life Function and Disability Instrument (LLFDI)

The LLFDI is a valid and reliable self-reported instrument of functional limitations and disability. Higher scores indicate better function or lower disability. MCID estimates for small and substantial changes in overall function, basic lower extremity function, and advanced lower extremity function are 2, 3, and 4 points, and 5, 6, and 9 points, respectively.³⁶

Short physical performance battery (SPPB)

The SPPB is a summed score of performance (range 0–12; higher scores = better function) on 3 timed tests: standing balance, 5-time chair stand, and usual walk speed over 4 meters.³⁷ We also considered continuous times for chair stand time and walk speed separately. The MCID is 0.5 for the SPPB and 0.1 m/s for gait speed.^38,39 Chair stand times ≥12 s predict a 2.4-fold increased risk of falls in older adults.⁴⁰

Stability

Postural sway

Participants were instructed to stand still with their feet together, hands on their hips, and eyes closed for 30s. Wireless Opal inertial sensors placed on the lumbar region and Mobility Lab software (version 2; Clario APDM Precision Motion, Portland, OR, USA) and algorithms were used to measure sway variables.^41–43 More sway is associated with higher fall risk.^42,44

Mobility

Timed-Up-and-Go (TUG)

TUG evaluates the time (s) it takes to rise from a chair, walk 7 meters at a fast pace, turn around, and return to a seated position.⁴⁵ Slower TUG times are associated with an increased risk of falls and disability.^46–48

Gait characteristics

In addition to the SPPB usual walk test, we measured gait characteristics during a 4 meter usual walk using an electronic gait mat (GAITRite^®, CIR Systems, Sparta, NJ, USA).

Statistical Analysis

Baseline group differences in demographics and outcomes were tested using two-sample t-tests for continuous variables and Chi-square tests of proportion or Fisher’s exact tests for categorical variables. Between-group changes in outcomes were reported as raw values (i.e., mean change in LIV group – mean change in UC group) with corresponding Cohen’s d effect sizes (using pooled standard deviation [SD]) with the effect-size package.⁴⁹ To estimate adjusted effect sizes, we used linear regression models with individual-level change scores (12 weeks – baseline) as the outcome while adjusting for age and cancer diagnosis (breast versus other). Cohen’s d effect sizes of .20, .50, and .80 were interpreted as small, medium, and large, respectively.⁵⁰ All analyses were conducted in R version 4.2.2.⁵¹

RESULTS

Participant characteristics

Thirty-eight participants were randomized to LIV (n=20) or UC (n=18; Figure 1; Table 1). Participants were primarily female (89%), white (89%), and non-Hispanic (95%), with a bachelor’s degree or higher (71%). The median age was 65 (IQR: 58–68) years, with a mean of 59.6 ± 9.1 years in LIV and 65.9 ± 9.8 years in UC (p=0.05). Most participants had breast cancer (63%). The median time since completing chemotherapy was 18 (IQR: 6–39) months, and the average baseline FACT/GOG-Ntx 4-item score was 5.61 ± 3.74. One participant in the UC group was excluded from efficacy analyses due to unexplained and implausible changes in neuropathy symptoms (i.e., >2 SD change).

Figure 1.

CONSORT diagram. LIV: low-intensity vibration, UC: usual care.

Table 1.

Demographic and clinical characteristics. Data are presented as mean (SD) or number (%).

Characteristic	Full sample n=38	LIV n=20	UC n=18	Range	p value
Age (years)	62.6 (9.9)	59.6 (9.1)	65.9 (9.8)	38.9–80.8	0.05
Female	34 (89%)	18 (90%)	16 (89%)		1.00
Race					0.36
White	34 (89%)	19 (95%)	15 (83%)
Non-white	4 (11%)	1 (0%)	3 (17%)
Hispanic	2 (5%)	0 (0%)	2 (11%)		0.40
Education – bachelor’s degree or higher	27 (71%)	16 (80%)	11 (61%)		0.16
Married/Partnered	24 (63%)	14 (70%)	10 (56%)		0.56
Cancer type					0.11*
Breast	24 (63%)	16 (80%)	8 (44%)
Colon	8 (21%)	3 (15%)	5 (28%)
Lung	3 (8%)	0 (0%)	3 (17%)
Lymphoma	1 (3%)	0 (0%)	1 (6%)
Pancreatic	1 (3%)	1 (5%)	0 (0%)
Uterine	1 (3%)	0 (0%)	1 (6%)
Time since diagnosis (months)	30.7 (19.5)	28.7 (15.4)	33.0 (23.4)	8–88	0.51
Time since chemotherapy completion (months)	23.1 (17.6)	22.0 (15.8)	24.3 (19.8)	3–60	0.69
Cancer stage at diagnosis					0.81
I	8 (21%)	4 (20%)	4 (22%)
II	19 (50%)	9 (45%)	10 (56%)
III	10 (26%)	6 (30%)	4 (22%)
Unknown	1 (3%)	1 (5%)	0 (0%)
Other cancer treatment
Surgery	23 (61%)	14 (70%)	9 (50%)		0.35
Radiation	13 (34%)	7 (35%)	6 (33%)		1.00
Hormone therapy	11 (29%)	8 (40%)	3 (17%)		0.22

LIV: low-intensity vibration, UC: usual care.

^*breast vs. all other cancer types.

Feasibility and safety

We enrolled 95% of our target sample (38/40) over 36 months, including a 16-month recruitment pause due to COVID-19 restrictions (Figure 1). Trial retention was 95% for the LIV group and 100% for the UC group. Some participants did not complete 12-week objective measures (n=2 LIV, n=4 UC) or PROs (n=1 UC) due to pandemic-related restrictions, illness, or refusal; however, we collected partial or complete 12-week data from 37/38 participants. Intervention retention was 80%. Adherence to twice-daily LIV sessions averaged 77.2 ± 18.2%, with 8 participants completing ≥80% of LIV sessions. Four participants who completed the study did not return tracking logs, but 2 provided verbal reports estimating ≥50% adherence. The other 2 participants provided no information about platform usage and were categorized as LIV non-completers.

Five participants in the LIV group reported 10 AEs that were determined to be study-related or possibly study-related, including muscle soreness, achiness, joint pain, bone pain, fatigue, and dysesthesia (all mild AEs). Dysesthesia led 1 participant to withdraw at 4 weeks and another participant to discontinue platform use at 8 weeks. There were no SAEs in LIV nor any AEs in UC.

Patient-reported symptoms

The LIV group reported greater improvements in FACT/GOG-Ntx 4-item score than UC (LIV: +1.4 ± 3.3 points, UC: +0.2 ± 2.8 points) with a medium adjusted effect size for change (Cohen’s d=.45; Table 2). The LIV group reported greater improvements in FACT/GOG-Ntx 11-item (LIV: +2.4 ± 5.0, UC: +0.8 ± 4.4) and TOI (LIV: +4.0 ± 12.4, UC: +0.1 ± 5.9) scores than UC, although the effect sizes were smaller (adjusted Cohen’s d=.33 and .35, respectively). There were negligible differences in global pain severity, global pain interference, and fatigue change between groups.

Table 2.

Baseline values and mean change in patient-reported symptom scores with unadjusted and adjusted differences between groups and effect sizes (Cohen’s d).

	LIV		UC		Unadjusted difference in change (95% CI)*	Unadjusted Cohen’s d	Adjusted difference in change (95% CI)**	Adjusted Cohen’s d**
	Baseline Mean (SD) n=20	Change Mean (SD) n=19	Baseline Mean (SD) n=17	Change Mean (SD) n=16
Neuropathy symptoms
FACT/GOG-Ntx (4-item)	5.2 (3.3)	1.4 (3.3)	6.3 (4.3)	0.2 (2.8)#	1.2 (−.9, 3.3)	.40	1.4 (−1.2, 4.0)	.45
FACT/GOG-Ntx (11-item)	24.4 (8.0)	2.4 (5.0)	27.8 (8.9)	0.8 (4.4)#	1.6 (−1.7, 4.9)	.34	1.5 (−2.5, 5.6)	.33
FACT/GOG-Ntx TOI	61.2 (18.2)	4.0 (12.4)	70.8 (14.7)§	0.1 (5.9)^	4.0 (−2.7, 10.6)	.39	3.5 (−5.3, 12.3)	.35
Pain
BPI-sf Pain Severity	2.8 (2.4)	−0.5 (2.2)	1.8 (2.1)	−0.4 (1.6)#	−.2 (−1.5, 1.2)	.08	0.2 (−1.4, 1.9)	.11
BPI-sf Pain Interference	2.8 (2.7)	−0.6 (2.3)	2.0 (2.9)	−0.2 (1.2)#	−.4 (−1.6, 0.9)	.20	−0.1 (−1.6, 1.5)	.04
Fatigue
PROMIS Fatigue SF8a	57.2 (9.9)	0.2 (6.4)	48.3 (12.0)	1.5 (8.7)	−1.3 (−6.7, 4.1)	.17	0.2 (−6.1, 6.5)	.03

BPI-sf: Brief Pain Inventory short form, FACT/GOG-Ntx: Functional Assessment of Cancer Therapy – Gynecologic Oncology Group Neurotoxicity, LIV: low-intensity vibration, PROMIS: Patient Reported Outcome Measurement Information System, TOI = trial outcome index, UC: usual care.

^*LIV - UC.

^**Adjusted for age and cancer diagnosis (breast vs. all other cancer types).

^§n=16,

^#n=15,

^{^}n=14.

Physical functioning

There was a large effect size for change in patient-reported basic lower extremity functioning (adjusted Cohen’s d=.80), with improved functioning with LIV (+5.3 ± 8.5 points) and slightly worsened functioning with UC (−0.7 ± 9.2 points; Table 3). Overall function, advanced lower extremity function, and disability scores showed small or negligible differences. There were small increases in total SPPB scores in both groups (LIV: +0.7 ± 1.1; UC: +0.9 ± 1.5), and small effect sizes for change in 5-time chair stand time and usual walk speed favoring LIV (adjusted Cohen’s d=.32 and .31, respectively).

Table 3.

Baseline values and mean change in physical functioning measures with unadjusted and adjusted differences between groups and effect sizes (Cohen’s d).

Physical functioning measure	LIV		UC		Unadjusted difference in change (95% CI)*	Unadjusted Cohen’s d	Adjusted difference in change (95% CI)**	Adjusted Cohen’s d**
	Baseline Mean (SD) n=20	Change Mean (SD) n=18	Baseline Mean (SD) n=17	Change Mean (SD) n=14
SPPB	10.6 (1.1)	0.7 (1.1)§	9.6 (2.0)	0.9 (1.5)^	−0.2 (−1.2, 0.8)	.16	−0.1 (−1.2, 1.0)	.08
5-time chair stand (s)	12.3 (2.6)	−1.3 (1.8)§	12.6 (3.7)#	−0.5 (3.1)α	−0.8 (−2.9, 1.4)	.31	−0.8 (−2.9, 1.4)	.32
Usual walk speed (m/s)	1.02 (0.27)	0.07 (0.14)§	0.95 (0.27)	0.05 (0.19)^	0.02 (−0.11, 0.15)	.15	0.05 (−.09, .19)	.31
LLFDI
Overall Function	62.8 (12.8)	1.34 (9.7)	63.0 (11.8)	2.1 (9.0)	−0.8 (−7.5, 6.0)	−.08	−1.2 (−9.2, 6.9)	.12
Basic Lower Extremity Function	70.5 (11.3)	5.3 (8.5)	76.9 (16.7)	−0.7 (9.2)	6.0 (−0.5, 12.5)	.68	7.0 (−0.5, 14.6)	.80
Advanced Lower Extremity Function	51.9 (10.7)	2.4 (9.0)	49.1 (13.2)	1.0 (4.5)	1.4 (−3.6, 6.5)	.19	2.8 (−3.5, 9.0)	.38
Disability	52.5 (6.5)	−0.7 (2.9)	56.7 (12.3)#	−0.2 (8.6)	−0.5 (−5.6, 4.5)	.09	−1.4 (−6.5, 3.8)	.23

LIV: low-intensity vibration, LLFDI: Late-Life Function and Disability Instrument, SPPB: short physical performance battery, UC: usual care.

^*LIV - UC.

^**Adjusted for age and cancer diagnosis (breast vs all other cancer types).

^§n=17,

^#n=16,

^{^}n=13,

^αn=12

Stability

There was a large effect size for one postural sway measure: sagittal centroidal frequency (adjusted Cohen’s d=.84) (Table 4), which decreased 0.15 ± 0.17 Hz (indicating improved stability) with LIV and slightly increased 0.01 ± 0.27 Hz with UC.

Table 4.

Baseline values and mean change in stability and mobility measures with unadjusted and adjusted differences between groups and effect sizes (Cohen’s d).

	LIV		UC		Unadjusted difference in change (95% CI)*	Unadjusted Cohen’s d	Adjusted difference in change (95% CI)**	Adjusted Cohen’s d**
	Baseline Mean (SD) n=19	Change Mean (SD) n=16	Baseline Mean (SD) n=16	Change Mean (SD) n=12
TUG (sec)	13.9 (3.7)	−1.9 (2.5)	14.5 (2.2)§	−0.7 (1.5)	−1.3 (−2.9, 0.3)	.59	−1.3 (−3.3, 0.7)	.60
Postural sway ***
Sway area (m2)	−1.76 (0.57)	−0.31 (0.66)§	−1.65 (0.90)	−0.48 (0.64)	0.17 (−0.35, 0.69)	.26	0.07 (−0.58, 0.71)	.10
Centroidal frequency (Hz)	0.07 (0.21)	−0.09 (0.16)§	0.06 (0.22)	0.01 (0.27)	−0.10 (0.29, −0.08)	.48	−0.14 (−0.35, 0.07)	.66
Coronal centroidal frequency (Hz)	−0.06 (0.19)	0.01 (0.20)§	−0.02 (0.15)	0.01 (0.19)	0.00 (−0.16, 0.15)	.01	0.03 (−0.16, 0.22)	.15
Sagittal centroidal frequency (Hz)	−0.13 (0.21)	−0.15 (0.17)§	−0.17 (0.18)	0.01 (0.27)	−0.16 (−0.36, −0.02)	.76	−0.19 (−0.41, 0.03)	.84
Jerk (m2/s5)	1.73 (0.90)	−0.54 (0.80)§	1.88 (1.01)	−0.33 (1.02)	−0.21 (−0.96, 0.54)	.23	−0.27 (−1.17, 0.63)	.30
Coronal jerk (m2/s5)	1.46 (0.83)	−0.46 (0.80)§	1.51 (1.23)	−0.28 (0.67)	−0.18 (−0.76, 0.40)	.25	−0.36 (−1.08, 0.36)	.49
Sagittal jerk (m2/s5)	1.75 (0.94)	−0.61 (0.79)§	1.88 (0.99)	−0.29 (1.16)	−0.33 (−1.15, 0.50)	.34	−0.29 (−1.26, 0.67)	.30
Mean velocity (m/s)	−1.40 (0.43)	−0.08 (0.77)§	−1.26 (0.64)	−0.31 (0.57)	0.23 (−0.30, −0.76)	.34	0.14 (−0.52, 0.80)	.21
Mean coronal velocity (m/s)	−2.16 (0.52)	−0.13 (0.64)§	−2.13 (0.53)	−0.24 (0.87)	0.11 (−0.51, 0.74)	.15	−0.02 (−0.71, 0.67)	.03
Mean sagittal velocity (m/s)	−1.79 (0.65)	0.06 (1.11)§	−1.62 (0.92)	−0.26 (1.03)	0.31 (−0.54, 1.17)	.29	0.27 (−0.80, 1.34)	.25
Gait characteristics
Speed (m/sec)	1.18 (0.30)	0.12 (0.13)§	1.14 (0.30)	0.08 (0.11)	0.04 (−0.06, 0.13)	.27	0.03 (−0.80, 0.14)	.26
Number of steps	6.4 (1.2)	−0.6 (0.6)§	6.3 (1.1)	−0.2 (0.8)	−0.4 (−1.0, 0.2)	.61	−0.4 (−1.0, 0.2)	.60
Cadence (steps/min)	110.4 (14.1)	4.0 (5.5)§	110.2 (17.1)	5.0 (5.6)	−1.0 (−5.4, 3.4)	.19	−1.0 (−6.1, 4.1)	.18
Step length (cm)	63.4 (10.2)	4.0 (5.2)§	61.3 (7.7)	2.1 (4.3)	1.9 (−1.8, 5.6)	.40	2.1 (−2.3, 6.5)	.43
Base of support (cm)	10.1 (3.2)	−1.3 (2.4)§	10.2 (3.3)	−0.1 (2.3)	−1.2 (−3.1, 0.7)	.51	−0.7 (−2.8, 1.4)	.29
Double-support time (% of gait cycle)	28.1 (4.0)	−0.7 (2.6)§	28.0 (4.1)	−1.2 (2.1)	0.5 (−1.3, 2.4)	.22	0.4 (−1.8, 2.7)	.17

LIV: low-intensity vibration, TUG: Timed-Up-and-Go, UC: usual care.

^*LIV - UC.

^**Adjusted for age and cancer diagnosis (breast vs. all other cancer types).

^***All postural sway measures are log transformed with base e.

^§n=15.

Mobility

There was a medium effect size for TUG time (adjusted Cohen’s d=.60), with a larger improvement (faster time) with LIV (−1.9 ± 2.5 s) compared to UC (−0.7 ± 1.5 s; Table 4). There was a medium effect size in the number of steps taken over 4 m (adjusted Cohen’s d=.60), with a larger reduction with LIV (−0.6 ± 0.6 steps) than UC (−0.2 ± 0.8 steps). There were small-to-medium effect sizes for other gait measures (Table 4).

DISCUSSION

These findings provide the first evidence of LIV’s feasibility and safety in cancer survivors with persistent CIPN, suggesting promising preliminary evidence of its potential salutary impact on key CIPN-related sequelae. CIPN remains an unmet clinical need, with evidence demonstrating associations with poor stability and mobility impairments; even mild CIPN symptoms are associated with clinically relevant deficits.^52–55 Our findings contribute to the growing evidence of vibration therapy’s effect on improved neuromuscular and musculoskeletal outcomes and symptoms across clinical populations, including cancer,^15,56–61 and add to the growing evidence demonstrating the safety of daily exposure to LIV across multiple patient cohorts.^{23,57,62–65} Collectively, our results suggest that a brief, daily dose of LIV is feasible, safe, well-tolerated, and may reduce symptoms and improve physical function, stability, mobility, and QOL in adult cancer survivors with persistent CIPN.

Despite the COVID-19 pandemic disruption, we enrolled 95% of our target sample and retained 97% of participants in the trial. Participants randomized to the LIV group completed >75% of LIV sessions , comparable to previously reported adherence rates that informed our hypothesis.^66,67 Together, the high adherence rates in the present study and prior LIV trials suggest that brief, daily LIV sessions at home are highly feasible, acceptable, and well tolerated by adult cancer survivors with CIPN. The convenience of home-based LIV therapy offers a significant practical advantage over higher magnitude modalities, which must be performed at a facility due to platform size and need for clinical supervision.

Five LIV participants reported mild musculoskeletal AEs consistent with commencing a new physical activity. Two participants who experienced dysesthesia withdrew from the study or prematurely discontinued the intervention. A sham-controlled trial is necessary to determine whether dysesthesia is exacerbated by LIV or by prolonged standing barefoot on the platform. We observed no patient-reported SAEs, further supporting LIV’s safety and aligning with previously published LIV interventional data.^{57,65,68–73}

LIV participants reported reduced numbness, tingling, and/or discomfort in the extremities after 12 weeks, as evidenced by a 1.4 (95% CI: −1.2, 4.0) point adjusted difference in change in 4-item FACT/GOG-Ntx scores between groups. This difference falls within the MCID range, reflecting a clinically meaningful improvement.²⁷ A larger controlled trial is needed to confirm these findings and the implications for clinical practice. FACT/GOG-Ntx scores were not significantly improved in previous trials of higher-magnitude vibration in cancer survivors.^74,75 In the present study, we observed a small effect size for change in 11-item FACT/GOG-Ntx scores between groups, consistent with LIV’s potentially smaller role in mitigating the functional consequences of CIPN.

LIV also improved objective measures of physical functioning, stability, and mobility. Patient-reported basic lower extremity function improved to a greater extent from LIV compared to UC, surpassing the MCID threshold, but will also need to be confirmed in larger trials.³⁶ Compared to UC, LIV also led to greater improvements in objective measures of mobility (i.e., faster TUG time; improved gait quality) and postural stability (i.e., decreased sagittal centroidal frequency), consistent with previous findings that LIV mitigated deteriorations in postural stability in a bed rest study in healthy adults.⁷⁶ Gait patterns were improved in the LIV group, evidenced by a greater reduction in step number during a 4 meter walk and longer step length compared to UC. This finding is consistent with previous reports of vibration interventions improving gait quality in older adults and people with diabetic neuropathy.^63,77,78 Older adults and individuals with neuropathy adopt a conservative gait pattern characterized by reduced speed, shorter and more frequent steps, and more time spent in the double-support phase of the gait cycle. A conservative gait pattern, slow TUG time, and increased postural sway are associated with greater falls and disability; thus, reversing these changes could prevent the long-term sequelae of CIPN.^53,79

Early signals of LIV’s efficacy are likely related to external vibration eliciting physiologic adaptations in the nervous and musculoskeletal systems.⁸⁰ LIV increased paraspinal musculature in women with osteoporosis, prevented neuromuscular junction degeneration in age-induced sarcopenia, and preserved muscle endurance and postural control in people confined to bed rest.^71,76,81 Preliminary findings also suggest that LIV elicits neuromuscular adaptations by stimulating muscle contractions through the tonic vibration reflex, increasing muscle activity and motor-unit recruitment, and likely stabilizing the joints.⁸² Future studies need to elucidate the mechanisms of LIV.

Strengths of this study include its randomized design, successful accrual (despite COVID-19), and focus on safety and feasibility as foundational endpoints to inform future investigation. For secondary outcomes, we utilized rigorous and evidence-based measures of CIPN-related clinical outcomes and objective measures of function, stability, and mobility. Limitations include this study’s homogenous patient population, reflecting the patient population at our single academic center. Due to limited study resources, outcome assessors were unblinded, but the use of sensors to measure study outcomes limited potential bias. Subjective data may be prone to bias; however, objective data are more immune to bias and align with the subjective improvements reported by the LIV group. Delivering a sham intervention for the control group would reduce potential subjective data bias in future trials.⁶⁶ We were unable to retrieve platform usage logs from all LIV participants and thus may have misestimated LIV adherence. Finally, we had limited available data about previous neurotoxic chemotherapy to evaluate whether the response to LIV differed by chemotherapeutic drug class or total chemotherapy exposures. Future studies might focus on CIPN related to more common adjuvant platinum- and/or taxane-based chemotherapies that have a similar etiology of CIPN.⁸³

As cancer survivors continue to live longer after receiving effective neurotoxic chemotherapy, the prevalence of CIPN will continue to increase. CIPN’s long-term sequelae of persistent pain, numbness, and instability lead to impaired physical function and mobility that limit cancer survivors’ independence while increasing the risks for falls and hospitalizations. Cancer survivors require effective, practical, and accessible interventions to mitigate CIPN sequelae and promote independence. As a critical first step, we have shown that LIV is feasible, safe, and well-tolerated among cancer survivors with persistent CIPN. The next steps include a double-blind, sham-controlled trial to evaluate LIV’s impact on CIPN-related clinical outcomes.

CONTEXT SUMMARY.

Key objective:

Is low-intensity vibration (LIV) therapy (where patients stand on a small platform that delivers a barely perceptible dose of vibration) a safe, feasible, and effective treatment for chemotherapy-induced peripheral neuropathy (CIPN) and associated deficits in physical functioning, mobility, and stability?

Knowledge generated:

LIV is safe (no serious study-related adverse events) and feasible (97% retention and 77% adherence to LIV sessions performed at home) in cancer survivors with persistent CIPN symptoms. After 12 weeks, LIV improved patient-reported CIPN symptoms and physical functioning, as well as objective measures of mobility and stability compared to a usual care group.

Relevance:

LIV shows promise as an effective, practical, and accessible treatment for mitigating CIPN and associated functional sequelae and improving quality of life in people living with CIPN. The results of this pilot study will need to be confirmed with a larger, sham-controlled phase III trial.