A Scoping Review of Chemotherapy Induced Peripheral Neuropathy-Related Gait Abnormalities in Children with Cancer

Abstract

Purpose:

Chemotherapy-induced peripheral neuropathy (CIPN) with associated weakness, areflexia, neuropathic pain, and sensory loss, is a common occurrence in children treated for cancer. However, accurate, quantifiable descriptions of gait deviations due to CIPN are lacking. This scoping review explores common gait abnormalities in children with CIPN.

Methods:

Using extensive database and manual literature reviews, 3 relevant studies were identified that assessed CIPN-related gait changes in children.

Results:

Though each study used different methods to analyze gait patterns, decreased ankle dorsiflexion with slower and shorter gait were common findings across the studies. Changes to ankle dorsiflexion range of motion can lead to slower walking speeds and compensatory changes that can cause developmental effects on a child’s gait.

Conclusions:

Relevant research on the topic of CIPN-related gait abnormality in children with cancer identifies knowledge gaps and need for more studies of the effects of CIPN on children.

Introduction

Muscle weakness, areflexia, neuropathic pain, and sensory loss are the hallmark signs of chemotherapy-induced peripheral neuropathy (CIPN) in children being treated for cancer.^1-7 These sensory and motor disruptions represent a side effect of several commonly used cytotoxic agents that leave clinicians and families with the difficult decision of choosing between reducing, or omitting, an effective agent or risking debilitating impairments that may follow a child well into adulthood. The short- and long-term neurotoxicity of vinca alkaloids, platinum compounds, etoposide, and others must be weighed against their potential benefit in cancer survival. Early detection of these problems can lead to successful mitigation of the more severe symptoms of CIPN; however, no universal gold standard for CIPN screening currently exists.^1,2,4,8 This is partially due to the varied presentation of CIPN across childhood cancer populations.^1,3,9 Even in the case of Acute Lymphoblastic Leukemia, where vincristine is the most common cause of CIPN, the presentation of neuropathy may affect sensory, motor, and autonomic divisions of the nervous system to varying degrees in different children.¹

The scales most widely used to grade CIPN include common toxicity criteria, functional assessment, quality of life assessment, and composite scales, none of which have been satisfactorily used for evaluation due to subjective symptoms such as numbness or fatigue.² The Common Terminology Criteria for Adverse Events (CTCAE) is commonly used to rate the severity of CIPN symptoms in clinical trials. The CTCAE defines peripheral motor neuropathy as: a disorder characterized by damage or dysfunction of the peripheral motor nerves and severity is assigned using a 5-point scale as follows. Grade 1: clinical or diagnostic observations only; Grade 2: moderate symptoms; limiting instrumental activities of daily living (IADL); Grade 3: severe symptoms; limiting self-care IADL; Grade 4: life-threatening consequences with urgent intervention indicated; and Grade 5: death.¹⁰ The CTCAE defines peripheral sensory neuropathy as: a disorder characterized by damage or dysfunction of the peripheral sensory nerves and similarly grades sensory neuropathy on the same 5-point scale.¹⁰ These measures poorly discern minute clinical changes, are subject to observer bias, and are coarse measures of changes affecting children. Cavaletti et al.² demonstrated a trained panel of oncologists could not consistently interpret the actual neurological impairment faced by people with CIPN. Furthermore, this measure has not been shown to distinguish statistically significant changes in CIPN.^3-5 Some have suggested electromyography or nerve conduction testing as a viable clinical assessment tool.^5-7 However, these studies are invasive, painful, and often clinically inconvenient and have thus failed to be widely adopted by clinicians treating the children.

Further complicating this picture is the inability of some children with cancer, especially younger or children with developmental delay, to describe the complex sensory, motor, or autonomic symptoms they are experiencing. Of the more widely accepted clinical CIPN assessment tools for use with the childrenvis the Pediatric-modified Total Neuropathy Score.¹¹ This tool, modified from the original work of Cornblath et al.¹¹ and validated by Gilchrist and Tanner, combines patient reporting of sensory and functional symptoms commonly associated with peripheral neuropathy with objective testing of sensation components.^11,12 Validation and use of this metric and others like it for younger populations have proven difficult due to the metric’s inherent dependence on the patient’s verbal communication. These metrics also do not account for normal, age-dependent, developmental variation in motor skills observed in young children. One study found that 23% of participants screened positive for a prior development vulnerability, such as delayed early motor milestones.¹³ These developmental vulnerabilities can alter the presentation of CIPN and should be taken into account for assessment of clinical changes of CIPN, especially for young children with delayed or in the early phases of motor learning.

Because of the broad effect of CIPN on motor performance and balance, several authors have noted gait disturbances in children with CIPN.^1,8,9 Questions about tripping or difficulty with walking are common subjective measures of function for this population. Indeed, a large number of children with an oncology diagnosis are already walking at time of diagnosis.^14,15 Objective measures of gait have been developed, including analysis systems with kinematic parameters measured through reflective markers attached to subjects and tracked using an optical motion measurement system to measure subjects’ gait deviation as compared to an average normal gait.¹⁶ Most of the gait analysis literature focuses on deviations associated with cerebral palsy, hereditary neuropathies, and post-stroke hemiplegia.¹⁷ Gait movements can be tested with cluster analysis which uses a motion capture system to allow a set of mathematical algorithms to identify gait patterns of homogenous groups (clusters) of disease conditions; however, there is no consensus about the best approach to this type of analysis.¹⁷ Because of this inconsistency with observational qualitative gait analysis for different pathologies, better recognition of the more common gait abnormalities stemming specifically from CIPN may help clinicians to identify and intervene earlier to remediate CIPN symptoms in children with cancer. The focus of this review is to compile data from the current literature identifying commonalities in what is known about CIPN-related gait patterns.

The following research question was developed for a scoping review: What common gait abnormalities are observed in children with chemotherapy-induced peripheral neuropathy?

Methods

A search strategy was developed based on the framework and stages proposed by Arksey and O’Malley and recommendations made by Levac and colleagues.^18,19 A search was executed using the following databases: Medline, CAB Abstracts, CINAHL (Cumulative Index of Nursing and Allied Health), PubMed central, Cochrane Library, and Web of Knowledge, Science Direct following input from a pediatric oncologist, a physical therapist in academia, a pediatric physical therapist specializing in pediatric cancer, and a university librarian. The following final search terms were included: (CIPN OR Chemotherapy-Induced Peripheral Neuropathy OR Chemotherapy Induced Peripheral Neuropathy OR VIPN OR Vincristine-Induced Peripheral Neuropathy OR Vincristine Induced Peripheral Neuropathy) AND (child OR children OR infant OR Toddler OR pediatric OR paediatric) AND (Gait Patterns OR Gait Characteristics OR Gait Deviations) AND Mobility AND Posture AND Pain AND Physical Activity AND Physical Functioning. The search was limited to articles published since 2000 since the primary tool for measuring peripheral neuropathy: the Total Neuropathy Score (TNS), was validated in 1999.¹¹ Pediatric versions of the TNS were not used until later.⁸ Articles were required to be in English for consideration. Additionally, some manual searching of journals and citations was used to identify additional key articles that may not have been included in database search results. This process took place on the websites for each of the databases searched. A reviewer read the titles of publications listed under the "similar articles" and "cited by" sections for any additional articles appropriate for further review.

Prior to screening and review, the authors established criteria for inclusion and exclusion of resulting articles. For inclusion in this review, studies must include (a) objective measurement of gait in a (b) pediatric cohort who have (c) undergone treatment for a pediatric malignancy and (d) using chemotherapy with known neurotoxic effects. Studies that did not meet the 4 criteria were excluded.

Results

Our initial search yielded 141,984 articles; 24 additional articles were identified through manual journal search and from citation review. After screening for duplicates, a total of 19,591 articles remained. From these, non-relevant or non-English language articles were removed by review of abstracts. The remaining 113 articles received a more thorough, full-text review. An additional 110 articles were excluded after full-text review due to lacking 1 or more of the previously mentioned criteria, such as gait measurement, pediatric cohort, treatment for pediatric malignancy, and/or using chemotherapy with known neurotoxic effects. The full-text review left 3 articles deemed to be relevant to the specific question based on this review’s criteria. (Figure 1).

Flowchart for scoping review article selection.

The patient population, analyses, and findings of the 3 studies are described in Table 1 and summarized as follows:

Table 1 –

Scoping Review Results Categorized by Participants and Gait Analysis.

Author (year)	Study Type	Participant Characteristics*	Timing of Gait Analysis	Type of Gait Analysis	Primary Results
Gilchrist et al. (2016) 20	Prospective Observation	N = 52 Sex: 30 females; 22 males Age: 11.3 ± 4.5 years Control: 52 (11.3 ± 4.5 years) Cancer Types: ALL (44%), Hodgkin or non-Hodgkin lymphoma (35%), non-CNS solid tumor (21%) Chemotherapy: Vincristine (all patients, mean cumulative dose of 17.5 ± 8.8 mg/m2), methotrexate (30 patients, mean cumulative dose of 132.3 mg), vinblastine (4 patients, mean cumulative dose of 18.8 mg/m2)	ALL patients tested: end of delayed intensification stage (~ 6 months into treatment) Hodgkin or non-Hodgkin lymphoma or other solid tumors tested: end of treatment (~3-6 months into treatment) Average time in treatment: 6.1 ± 2.4 months Other: Each patient’s gait was tested before and after a 6 Minute Walk Test (6MWT)	Instrumented Gait Analysis: (GaitRite electronic walkway) velocity (cm/s) cadence (steps/min) step length (cm) base of support (cm) time in double support (percent of the cycle) percent of time each portion of the foot is in contact with the pressure mat	Slower walking velocity in oncology participant compared to healthy children due to a decreased step length and wider base of support Tightness in ankle dorsiflexion helps explain variances
Beulertz et al. (2016) 21	Cross-sectional	N = 13 Sex: 8 females; 5 males Age: 11.14 ± 3.53 years Control: 13 (11.29 ± 3.42 years) Cancer Types: ALL (30.8%), Hodgkin or non-Hodgkin lymphoma (15.4%), CNS tumor (23.1%), non-CNS solid tumor (30.8%) Chemotherapy: 10 of out 11 participants that utilized chemotherapy agents had received vincristine during medical treatment (doses not specified)	After completion of chemotherapy Mean time since cessation of medical Treatment = 1.37 ± 1.07 years Mean age at diagnosis 9.30 ± 4.20 years	Instrumented Gait Analysis: Microgate Optogait 2D Gait Analysis System: length parameters in centimeters (step length, stride length, step width) gait cycle parameters in percentage (stance phase, swing phase, single-limb support, double-limb support, loading response, and pre swing phase) time parameters in seconds (step time and gait cycle), and cadence (steps per min) and walking speed in meters per second.	Decreased dorsiflexion range of motion in oncology sample compared to healthy children in both knee flexed and extended positions Significant differences in stance, swing, and preswing phase
Wright et al. (2017) 22	Cross-sectional	N = 17 Sex: 10 females; 7 males Age: 11.2 ± 5.7 years Control: 10 (12.4 ± 5.5 years) Cancer Types: ALL Chemotherapy: Vincristine (doses 1.5 mg/m2 weekly during induction and 2.0 mg/m2 every three weeks during consolidation and maintenance therapy)	7 subjects were tested during continuation treatment 10 subjects were tested after completion of treatment Average time off treatment: 35.7 ± 28.5 months	Instrumented: 3-DMA: VICON 8 MX-40 camera system (Vicon Motion Systems Inc.) Gait: Plug-in-Gait 16 marker set (2.0.1), Polygon software (Version 3.1) and Plug-in-Gait 16 marker set (2.0.1), Polygon software (Version 3.1), and three 1000 Hz force platforms Advanced Mechanical Technology Inc. Video Analysis: 2 synchronized video cameras Electromyography (EMG): surface, gastrocnemius and tibialis anterior muscles	significantly less dorsiflexion at initial contact significantly shorter step lengths and slower velocities compared to the control group

ALL – acute lymphoblastic leukemia CNS – central nervous system

^*The demographic data in the table was reformatted to allow for easier comparisons including converting percentages to numeric values. None of the information was altered in a way as to change the data from the studies included.

Gilchrist et al. evaluated 52 children with age-matched controls in a prospective fashion using instrumented quantitative gait analysis. Major findings were that compared with controls, children walking at a self-selected comfortable pace had significantly slower normalized velocity, decreased step length, wider base of support, decreased distance on 6 Minute Walk Test (6MWT), decreased cadence after the 6MWT, increased forefoot contact time after 6MWT. Tightness in ankle dorsiflexion helps explain these variances.²⁰

Beulertz et al. evaluated 13 children with age- and sex-matched controls in a cross-sectional study. They concluded that after completing all medical treatment (including chemotherapy), significant differences exist between the oncology sample and the healthy comparison sample in the following ways: (1) stance phase was significantly increased, (2) swing phase was significantly decreased, and (3) pre-swing phase was significantly increased in the cancer group compared to the control group. The researchers found decreased active dorsiflexion range of motion in oncology samples compared to children without cancer in both knee flexed and extended positions.²¹

Wright et al. evaluated 17 children with 12 controls in a cross-sectional sample of children with cancer. Their gait analyses included video, electromyography (EMG), and software analyses. Based on this sample, they reached the following conclusion that those with CIPN exhibited a spectrum of EMG, kinematic, kinetic, and temporal spatial deviations through the gait cycle without a single set of motion analysis findings to describe gait in participants. The researchers found significantly less dorsiflexion at initial contact as well as significantly shorter step lengths and slower velocities compared to the control group.²²

Discussion

There are relatively few examples in literature that define the gait abnormalities related to CIPN in children. Despite the limited studies, this review highlights several common gait abnormalities and methods of detection and assessment. All 3 studies found significant differences in ankle dorsiflexion with an increase in tightness and/or a limitation of ankle range of motion due to abnormalities of the tightness of the tibialis anterior and hypoextensibility of the gastrocnemius muscles resulting in decreased power generation of the step-off. A common trend towards slower walking speeds with significantly slower normalized velocity, decreased step length, decreased swing phase, and increased pre-swing with reduced ankle power is noticed in the 3 studies.

Given that the effects of CIPN are noted first and more significantly at the distal lower extremity, the ankle and knee are likely more pertinent in assessing CIPN-related gait effects than more proximal joints. Of note, the ankle and knee joints are easier and more reliable to measure compared to the trunk, hips, and pelvis.²³ The tightness in ankle dorsiflexion or reduced range of motion of the ankle was found in each of the reviewed studies. Changes in the joint angle measurements are indicative of diseases or injury as opposed to immaturity in development.²⁴ A degree of ankle dorsiflexion in the stance and swing phase, allowing for adequate foot clearance, is required for fluid walking strides. The researchers found that balance scale scores and tightness in ankle dorsiflexion best explained the variance between CIPN participants and controls.²⁰ Therefore, dorsiflexion of the ankle could potentially be measured manually, resulting in a simplified means of assessing gait deviations through standardized methods.

Gilchrist et al. found no significant difference in cadence between children and controls.²⁰ To accommodate reduced ankle range of motion, the CIPN subjects reduce step length, effectively prompting an earlier heel rise that compensates for reduced ankle range of motion and knee strength.²⁰ While the participants with peripheral neuropathy were able to increase their walking speed when asked, they still walked slower than their age- and sex-matched peers on every gait assessment.²⁰ The CIPN group in the Wright et al. study also had significantly shorter step lengths and slower velocities compared to the control group.²² Beulertz et al. similarly found significantly reduced walking efficiency in the oncology sample with the 6MWT test.²¹

Limited dorsiflexion may cause compensatory movements such as knee hyperextension or early heel rise. Both ankle range of motion and knee extension strength are important functions for normal gait and managing stairs.²⁵ In the Beulertz et al. study, significant differences in ankle dorsiflexion range of motion in both knee flexed and knee extended positions were not limited to children with acute lymphocytic leukemia (ALL) or those treated with vincristine, and were seen in all pediatric cancer diagnoses included in their study.²¹ The presence of these deficits in children not likely to be exposed to neurotoxic chemotherapy suggests that other treatment-related factors, such as immobility, may play a role in observed gait deviations in childhood cancer survivors.²¹ Wright et al. showed significantly less dorsiflexion at initial contact during the stride and a reduction in knee flexion at loading associated with the decreased dorsiflexion in children undergoing treatment for ALL.²² The pre-swing was characterized by less hip extension, decreased, and delayed peak ankle plantar flexion, and reduced ankle movement and power.²² Participants with peripheral neuropathy had significantly less peak dorsiflexion in swing; however, foot clearance was still achieved in all but one participant.²²

One of the major findings from Gilchrist et al. was an increased forefoot contact time after 6MWT, commonly known as foot-slap.²⁰ This change in forefoot contact may reflect decreased eccentric control and neuromuscular fatigue of weakened ankle dorsiflexors in children with vincristine-related neuropathy. Gilchrist et al. also found that children with CIPN increased the percentage of the step with the forefoot in contact with the ground but maintained the same percentage of the step with the heel in contact with the ground as the control participants.²⁰ This increase in forefoot contact may reflect the slow lengthening muscle contraction control and fatigue from the weakened ankle dorsiflexors due to CIPN.

Limited range of motion in ankle dorsiflexion causes compensatory movements listed previously and creates an abnormal gait pattern called steppage gait. Foot-drop, or steppage gait, is a neuromuscular gait abnormality characterized by the patient lifting the swinging leg higher than normal to compensate for decreased ankle dorsiflexion causing the toes to have a decreased clearance of the ground.²⁶ While foot drop can be accurately measured with electronic quantitative gait analysis, steppage gait can also be consistently identified by the trained eye of providers familiar with CIPN.

Certain interventions have been shown to restore dorsiflexion and gait range of motion, thereby potentially reducing long-lasting detrimental effects of CIPN on the gait of developing children. Means of increasing the strength of dorsiflexion and increase in passive dorsiflexion range of motion through orthotics use can be a solution to some of the deficits seen in CIPN.²⁷ Manual stretching by a trained physical therapist has been shown to help guide children to achieve proper body alignment and increase ankle dorsiflexion active range of motion.²⁵ Without these interventions, children with CIPN may compensate with other movements that can damage the integrity of the ankle or knee causing long-term effects on their joints and gait. The availability of effective interventions emphasizes the importance of accurate analysis and identification of CIPN in children in order to intervene before deficits in gait become severe and more difficult to treat.

In general, qualitative gait analysis can be a particularly helpful, readily available clinical tool when assessing gait. However, when considering the wide variability in the communication skills of children with cancer, many younger children and children are developmentally unable to communicate about complex CIPN symptoms.²⁸ Preverbal children may benefit from quantitative gait analysis that can identify these common patterns and are not dependent upon a patient’s ability for verbal communication. However, access to quantitative gait analysis labs is limited in most areas of the United States, a measure to assess changes to deviations from an average normal gait need to be obtainable without the high-technology process. Previous studies have shown that pain behaviors such as crying, facial expressions, and physiological signs such as shivering, sweating, and breath holding are consistently and accurately reported for pain intensity scales for children who lack the ability to self-report symptoms.²⁹ By using metrics that minimize the need for the patient’s verbal input, CIPN assessment tools can be used on a broader population.

These studies provide direction for evaluation of gait in children at risk for CIPN. The commonalities in gait-related impairments and compensation strategies reported by multiple authors and uncovered in this review lend credibility to the use of qualitative observational gait analysis in detecting and grading severity of CIPN. However, the small number of studies with limited number of children highlights a knowledge gap that warrants additional study.

This scoping review was subject to several important constraints which may have inherently limited the results. First, without translation services available, only articles written or translated into English were reviewed within the databases. This potentially excludes any non-English articles through language bias. Secondly, each database search was limited to articles published since 2000 due to lack of validated tools for neuropathy assessment prior to 1999. Lastly, as this review is scoping in nature, no thorough evaluation of the quality of each of the studies performed was included. Results of such a review are intended to present the breadth of information on a given topic and caution should be exercised when interpreting the quality of the individual articles contained within.

Conclusions

Even with the limited number of studies presented in this scoping review, there is a clear need for clinicians to assess the degree of CIPN in children receiving neurotoxic chemotherapy treatments. The gait and motor development anomalies that result from unaddressed CIPN in developing children can have long-lasting effects, as children could adapt to these deviations and carry them forward through the remainder of their gross motor development. Additionally, these studies primarily included children with ALL. While ALL is the most common childhood cancer associated with a high risk of CIPN, the effects of CIPN associated with different treatment combinations used in other childhood cancers should be assessed as well. The lack of quality data results in the absence of accepted standards for monitoring changes in gait deviations and early identification of clinical features of CIPN. Delayed diagnosis of CIPN often leads to missing the window of early intervention to change the course of CIPN before more significant impairments develop. The authors believe that a simplified metric for assessing gait that does not involve costly, sophisticated qualitative observational gait assessment tools is needed to evaluate CIPN in clinical practice. Unique challenges exist in assessing CIPN in pediatric practice to allow for assessment of children of all abilities and developmental stages. Because many children with cancer are preverbal or nonverbal, assessment tools that rely on patient communication are not highly reliable. Assessment of CIPN in children must account for the patient’s stage in motor development and not rely on the patient’s verbal input, but rather on assessment of their movements during play in a comfortable environment. Tools that can assess CIPN in a clinical setting with or without a quantitative gait analysis lab, therefore, would be beneficial for clinicians to assess CIPN in real time in order to adjust neurotoxic drug dosing and/or guide therapy plans to reduce the long-term effects of CIPN in children.

WHAT THIS EVIDENCE ADDS.

Current evidence

Chemotherapy Induced Peripheral Neuropathy (CIPN) is a common complication of childhood cancer treatment and can significantly disrupt gait mechanics.^{1, 12, 13} There is no consensus about how to best analyze the motor effects of CIPN, particularly in developing children. The Pediatric-modified Total Neuropathy Score is commonly used, but does not account for normal age-dependent, developmental variation in young children.⁹ A strategic search of Medline, CAB Abstracts, CINAHL, PubMed central, Cochrane Library, Web of Knowledge, and Science Direct databases yielded numerous background sources, with most published after 1999.

Gap in the Evidence:

In order to investigate how CIPN affects gait in children, the following question was developed for this scoping review: What common gait abnormalities are observed in children with CIPN?

How does this study fill this evidence gap?:

We compiled data to identify both commonalities and knowledge gaps in the field of CIPN-related gait patterns. Ultimately, 3 studies were identified as relevant to our question.^{20, 21, 22} These studies suggest that the effects of CIPN primarily start with loss of ankle dorsiflexion which leads to more proximal gait abnormalities. This also identified a need for further studies of CIPN, particularly in young children.

Implication of all the evidence:

Though only 3 studies met criteria for our scoping review, they demonstrate a clear impact of CIPN on gait patterns in children. Therefore, assessment of gait in children is important in assessing short- and long-term effects of CIPN in children exposed to neurotoxic chemotherapy. New metrics need to be developed to assess gait alterations in children with cancer that also account for the child’s stage in motor development. These metrics would be beneficial for clinicians to assess CIPN in real-time and adjust therapy accordingly to reduce the long-term effects of CIPN in children.