Children and adolescents who have undergone treatment for cancer are at risk for a number of adverse health effects, both short- and long-term, which have the potential to impact physical function.1 One of the most common side effects of cancer treatment in both pediatric and adult populations is chemotherapy-induced peripheral neuropathy (CIPN).2–4 Some of the most commonly used neurotoxic agents in pediatric oncology practice include vincristine, cisplatin, and etoposide.5 Vincristine is used widely in pediatric oncology for treatment of cancers such as leukemia, lymphoma, Wilms tumor, and neuroblastoma, along with many others. Vincristine is thought to bind to microtubules and thus cause axonal transport deficits in peripheral axons, resulting in CIPN.6 Other chemotherapeutic agents can cause peripheral nerve dysfunction through damage to sensory neuron cell bodies, mitochondrial dysfunction, and demyelination.6 In children and adolescents, CIPN can result in mild to moderate deficits in distal strength, sensory impairments, impaired deep tendon reflexes, and neuropathic pain syndromes.5 While the incidence of CIPN has been documented both during and after cancer treatment,7–10 the effect of CIPN on motor performance such as balance control has received limited attention in children and adolescents.
To have optimal balance control, multiple body systems need to be functional. Appropriate visual, somatosensory, and vestibular input needs to be transmitted to the central nervous system (CNS). Integration of sensory information at the level of the CNS is necessary to provide feedback or create feed-forward strategies to maintain balance. Intact lower motor neuron, muscle, and joint function are required to implement the motor commands.11 Other factors, such as environment and cognition, may also have roles in balance control.11 The majority of children and adolescents receiving vincristine develop a mild to moderate neuropathy7,8 that can impact both the somatosensory input and the lower motor neuron output required for balance control. The Pediatric Modified Total Neuropathy Scale (ped-mTNS), a validated measure of CIPN,12 was found to be in the abnormal range for 86.5% during treatment and impairments occurred in both somatosensation and distal strength.7 By 6 months posttreatment, only 11.5% of subjects treated for acute lymphoblastic leukemia (ALL) continued to have abnormal CIPN scores, but 57% of those treated for lymphoma and 60% treated for solid tumors, such as Wilms tumor and rhabdomyosarcoma, continued to score in the abnormal range. Smith et al8 found similar results with on-treatment CIPN incidence using a different version of the Total Neuropathy Scale. CIPN persistence at least 1 year off of treatment has been documented in 16% to 34% of children treated for ALL.10,13,14 Thus, CIPN is a common and sometimes persistent issue for children undergoing neurotoxic chemotherapy.
Balance control may be impacted by CIPN. Decreased gross and fine motor performance in children treated for ALL has been identified.10,15–17 Wright et al17 tested 99 subjects who had completed treatment at least 1 year prior for childhood ALL and found that their balance control was poorer than their peers. Others have reported decreased balance control in survivors of childhood cancer, mostly in pediatric ALL, with some variability in incidence likely due to treatments received, time since treatment, and balance assessments.18 Gilchrist et al7 found that CIPN as measured by the ped-mTNS was associated with worse scores on tests of balance (rs = −0.626, P < .001) and manual dexterity (rs = −0.461, P < .001) while children were on treatment with vincristine for non-CNS cancers. Studies of adult survivors of pediatric cancers demonstrate that balance and functional mobility impairments are present in a greater percentage of ALL survivors than sibling controls.9,19
While CIPN and motor performance deficits have been demonstrated in children treated with neurotoxic treatment, few studies are available to investigate the relationship between the development of CIPN and motor impairments as well as their recovery after cessation of treatment. Previously, we reported on the incidence and short-term recovery of CIPN in a population of subjects being treated for non-CNS cancers with vincristine,7 and in this study we detail this same population's recovery in balance control. Thus, the aims of this study are to (1) describe the on-treatment incidence and recovery of balance impairment in a group of children receiving treatment for multiple non-CNS cancers with vincristine therapy; (2) investigate the association between balance impairment with disease and treatment factors; and (3) explore the association between CIPN and balance impairment.
Children and adolescents with a new diagnosis of ALL, lymphoma (Hodgkin and non-Hodgkin), or other solid tumors except for brain tumors, ages 5 to 18 years, who were patients of a children's oncology clinic between October 2009 and June 2014 were approached for participation. The study closed recruitment for children with ALL in August 2012 because the incidence of ALL is higher than the incidence of other childhood cancers. This was done to balance the number of children with ALL to the other diagnostic groups. All recruitment material, consent forms, and testing procedures were reviewed by the Institutional Review Board at the participating institution.
Demographic and Treatment Data
Parents of participants completed a questionnaire to provide demographic and medical history data. Information regarding cancer diagnosis and treatment was obtained from medical records by a trained abstractor and included diagnosis and staging, surgical interventions, cumulative dose of chemotherapeutic agents used both prior to initial testing and through the end of therapy, presence of dose reductions for neurotoxic side effects, duration of cancer treatments, and comorbid conditions.
Timing of Measures
Measures were completed for children undergoing cancer treatment a minimum of 2 months after the initiation of chemotherapy. In children with ALL, patients were measured within 2 weeks of the end of delayed intensification, which is approximately 6 months into treatment. In children with lymphoma and solid tumors, patients were measured 3 months after the initiation of chemotherapy. Subjects were tracked throughout treatment and tested again both at 3 and 6 months postcompletion of chemotherapy. Treatment for childhood ALL typically lasts 2 to 3 years while for most types of lymphoma and other solid tumor treatments are completed within 3 to 9 months.
The balance subtest of the Bruininks-Oseretsky Test of Motor Proficiency version 2 (BOT-2) was administered according to standard procedures.20 Raw scores were converted to scaled scores based on age and gender-matched normative data. Lower scores indicate increased impairment. A score of 15 is the normative mean and scores of 10 or lower are 1 standard deviation (SD) or more below the population norm. Neuropathy was evaluated using the ped-mTNS.12 This validated instrument captures sensory, motor, and autonomic symptoms; light touch, pin, and vibration perception; muscle strength of distal musculature; and deep tendon reflexes. The total score is based on the sum of items; higher scores indicate increasing symptoms and deficits. A score of more than 4 on the total scale score is considered abnormal. A sensory subscore is calculated as the sum of sensory symptom scores and scores from clinical examination of light touch sensitivity, pin, and vibration perception, with a maximum of 16 points. A motor subscore is calculated by adding scores from the subjective question on functional mobility and clinical examination of the strength item, for a maximum of 8 points. Data on the ped-mTNS scores over time were previously published,7 and are presented here as they relate to recovery of balance control.
Descriptive statistics were used to summarize the demographic and treatment characteristics of the study population. Balance and CIPN scores from the entire group were compared across time with repeated-measures analysis of variance (ANOVA). Mixed between-within ANOVA evaluated the diagnosis effect on balance. Means and standard errors for CIPN were compared between diagnostic groups at each time point with ANOVA as the general linear model for time × diagnosis for CIPN scores violated assumptions of variability of error and covariance. Spearman correlations were used to evaluate possible risk factors for association with balance scores. All analyses were done with SPSS (version 23, IBM, Armonk, New York).
The sample recruitment process is detailed in Figure 1. Overall, 112 patients met inclusion criteria and 86 (76.8% of eligible) completed initial measures. Sixty-five completed 6-month follow-up measures, representing 58.0% of eligible participants and 75.6% of those completing initial measures. The final sample of 65 was not significantly different from the sample meeting inclusion criteria based on age, gender, or diagnostic grouping.
Demographic and treatment information for the group as a whole and for different diagnostic categories is shown in Table 1. The total population had a mean age of 11.3 (4.2) years, ranging from 5 to 18 years of age at the first measurement time point, and 46.0% were male. The mean body mass index (BMI) percentile of the total group was 61.6 (29.1). Children treated for ALL were younger, more likely to be female, and had been treated for longer than the other diagnostic groups at the time of initial measurement. All participants were exposed to vincristine. Nine of 65 (14%) had vincristine dose reductions due to neurotoxicity per medical treatment study protocols or physician discretion at the time of initial measure and that increased to 26% by the time treatment ended. No participants received platinum-based chemotherapy and only 2 received bortezomib. None of the participants received cranial radiation. Most subjects received physical therapy (PT) during their cancer treatment, with 94% receiving an assessment and 82% receiving an intervention. Only 4 patients, all with lymphoma, did not receive a PT assessment. For those who did receive PT intervention at the institution (n = 51), their mean number of PT sessions was 14.7 (±13.8, range 1-51) and the number of units reported for balance training was on average 8.9 (±8.5, range 1-33). Two individuals were referred to other institutions for PT.
Balance scores as measured by the BOT-2 for the overall group (Wilks λ = 0.68, F = 11.75, P < .005; partial η2 = 0.32) improved over time (Figure 2). While on treatment, 78% of the participants (51/65) scored 1 SD or greater below the mean, with 29% (19/65) scoring more than 2 SD below the mean. This improved by 6 months posttreatment, with 53% of the participants at 1 SD or greater below the mean and only 6% (4/65) scoring 2 SD or more below the mean. While each of the treatment groups improved over time, the ALL group performed significantly worse than the other solid tumor group initially and at 6 months (F = 4.07; P = .02, partial η2 = 0.145). Although all groups improved balance following treatment, mean scores for all groups at all times remained below population norms mean scale score (15 ± 4) based on age and gender.20
Balance Recovery and Treatment Factors
Balance scores at 6 months off treatment were not significantly correlated with age at diagnosis or total cumulative doses of vincristine, intrathecal methotrexate, or etoposide, but an association between balance at 6 months posttreatment and BMI percentile was seen (rS = −0.30, P = .016). On-treatment balance scale scores were associated with both the number of PT sessions received (rS = −0.34, P = .015) and the number of balance training units charged (rS = −0.34, P = .014), such that the worse balance (lower) scores were associated with increased intervention. There was no significant association between PT sessions or balance units charged and either 6-month balance scores or change in balance scores between on-treatment and 6-month follow-up.
CIPN Recovery and Association With Balance Recovery
While improvement in balance was occurring during recovery postchemotherapy, neuropathy scores on the ped-mTNS were also improving (Wilks λ = 0.411, F = 35.1, P < .005, partial η2 = 0.59) (Table 2). When measured on treatment, 86% of the total group scored greater that 4 on the ped-mTNS indicating neuropathy. This declined over time posttreatment, and by 6 months after cessation of chemotherapy, only 42% of participants scored greater than 4. Neuropathy scores for the lymphoma group were significantly higher (worse) than those for the ALL group at 3 months posttherapy (F = 10.1, P < .005) while those in the ALL group had significantly lower scores (better) than those in both lymphoma and other solid tumor groups at 6 months (F = 8.0, P = .001) (Table 2). While the subjects were on-treatment, there was a moderate, negative association between balance scores and ped-mTNS scores (rS = −0.34, P = .005). This association was also seen between the 6-month off treatment balance and ped-mTNS scores (rS = −0.31, P = .01). Changes in ped-mTNS and BOT-2 balance scores from on-treatment to 6 months posttreatment were also moderately, negatively associated (rS = −0.33, P = .01).
Recovery of neuropathy occurred across all items and areas on the ped-mTNS, but at 6 months posttreatment, many subjects continued to have impairments in distal strength, decreased deep tendon reflexes, and sensory deficits (Table 3). While on-treatment, balance scores were more closely associated with motor composite item scores than with sensory composite item scores (rS = −0.39, P = .001 vs rS = −0.23, P = .065), the 6-month off treatment balance scores were more closely associated with sensory composite item scores on the ped-mTNS (rS = −0.35, P = .004) than with motor composite item scores (rS = −0.08, P = .541). When balance scores at 6 months were compared with individual items on the ped-mTNS, only the items of sensory symptoms (patient report of numbness, paresthesias, or distal pain) (rS = −0.29, P = .02) and clinical examination of vibration sensation (rS = 0.26, P = .04) were significantly associated.
In this longitudinal study of children and adolescents treated for non-CNS cancer, balance deficits found while patients were on-treatment improved, but more than half of the participants continued to have at least mild balance impairment by 6 months posttreatment. The improvement in mean balance control seen in all groups (ranging from 1.8 to 4.4) was above the reported minimally important difference (0.57) and the minimal detectable change (1.14) for the BOT-2 balance subscale, as reported by Wuang and Su.21 While children were in treatment for non-CNS cancers, we found a moderate association between balance scores on the BOT-2 balance subscale and CIPN as measured by the ped-mTNS, as we have previously reported.12 Although neuropathy improved significantly over time after cessation of chemotherapy, balance improved but did not fully recover. This was particularly striking in participants treated for ALL, as their neuropathy scores improved to near normal by 6 months, with a mean score of 2.4 (1.7), yet this population had the lowest overall balance scores. Treatment for ALL is significantly longer, an average of 2 to 3 years as compared to 3 to 9 months for the other cancer types, and the prolonged exposure to neurotoxic treatments may be a factor in the residual balance impairment.
Our finding that 78% of the children on-treatment for cancer have below average (greater than 1 SD below population means) balance and 29% score well below average (greater than 2 SD below population means) is consistent with the previous report of Reinders-Messelink et al,16 who found impaired balance in a small cohort of patients being treated for ALL. Interestingly, they found that 50% of their participants had impaired balance even before the initiation of chemotherapy potentially stemming from the effect of disease or medical interventions given during the diagnostic phase. In survivors of childhood cancer, it has been reported that 7% to 65% have balance impairment, but most of these studies have been completed in participants who are 1 year or more off treatment.18 In one study that assessed a group of participants treated for ALL between 0 and 12 months posttreatment, 10% had a balance impairment measured with the Movement Assessment Battery for Children–Edition 2.15 This is less than our finding that at 6 months posttreatment 53% of children and adolescents scored greater than 1 SD below mean on the BOT-2, but is close to our finding that 6% scored greater than 2 SD below mean. Our findings also extend these previous results in that they indicate that balance control is a potential issue not only for those treated for ALL but also for lymphoma or other non-CNS cancers.
The association between the mild to moderate balance control impairment seen in this population and the mild to moderate neuropathy that can occur with cancer and its treatment fluctuates over time. The moderate correlations between balance and neuropathy are likely due to multiple factors, and while peripheral nerve function is important, balance control is reliant on a number of factors outside of somatosensory input and motor output. There are aspects of nervous system function, important in balance control, which were not assessed here—vestibular function and central processing speed. Intrathecal administration of methotrexate or other agents may lead to CNS damage impairing balance,22 but we did not observe an association with intrathecal methotrexate dosage. In addition, the loss of ankle range of motion commonly seen in children treated for cancer may also be a significant factor that was not investigated in our study.23,24 Increased BMI occurs in pediatric cancer treatment, and our finding of an association between increasing BMI and decreasing balance control was also reported by Wright et al.17 Balance was more closely associated with motor signs and symptoms during treatment, but with sensory signs and symptoms after treatment and may in part reflect the greater improvement seen in muscle strength as compared with sensory signs over time as well as a potential focus on strengthening as a part of PT (Table 3).
It is concerning that balance scores did not return to normal for many patients by 6 months posttreatment, despite improved neuropathy and intervention by PT for 86% of the population. While the level of balance impairment is mild for most of the participants, when added to other seemingly minor deficits such as decreased range of motion, strength, and overall physical activity, it could lead to survivors with poorer overall health and early frailty. Ness et al9 tested adult survivors of childhood ALL who had a mean age of 35, and found that their balance scores on computerized dynamic posturography were comparable to normal 60- to 69-year-olds. Ness et al25 demonstrated early physiologic frailty in a large cohort of young adult childhood cancer survivors. If the decreased balance seen in our cohort does not continue to improve to normal levels, and instead plateaus, it could be one indicator of early frailty, lead to greater risk of falls, and poorer overall health.
There are limitations to this study. First, we removed participants who were lost to follow-up between the initial and final measures. While it is reassuring that the final sample did not significantly differ in age, diagnosis, or gender from those eligible for the study, there could have been other differences in those completing measures that could have affected the findings. While our measure of balance is a common standardized test used in PT, a test that captured higher level dynamic balance, might have been more sensitive in picking up the mild balance deficits typical in this population.
This longitudinal study described the incomplete balance recovery seen in children and adolescents who have been through neurotoxic treatment for cancer. While balance impairments occurring during cancer treatment improve over time, mean scores remain below population norms at 6 months posttreatment for multiple forms of cancer. Neuropathy also improves with time off treatment and is associated with balance impairment, but even those participants who recovered from CIPN continue to have balance impairments. Thus, clinicians should screen both pediatric and adult survivors of childhood cancer for balance impairments, which may persist both in those with residual CIPN and in those where it has resolved.
The authors would like to thank Katherine Wacker, Jessica Ovans, and Jeffrey Mason for their assistance with research testing and manuscript editing, Dr David Chapman for his helpful comments on the manuscript, and Francesca Schirber and Ann Logelin for medical records abstraction.
1. Wilson CL, Gawade PL, Ness KK. Impairments that influence physical function among survivors of childhood cancer
. Children (Basel). 2015;2:1–36.
2. Kandula T, Park SB, Cohn RJ, Krishnan AV, Farrar MA. Pediatric chemotherapy induced peripheral neuropathy: a systematic review of current knowledge. Cancer
Treat Rev. 2016;50:118–128.
3. Brewer JR, Morrison G, Dolan ME, Fleming GF. Chemotherapy-induced peripheral neuropathy: current status and progress. Gynecol Oncol. 2016;140:176–183.
4. van de Velde ME, Kaspers GL, Abbink FCH, et al Vincristine-induced peripheral neuropathy in children with cancer
: a systematic review. Crit Rev Oncol Hematol. 2017;114:114–130.
5. Gilchrist L. Chemotherapy-induced peripheral neuropathy in pediatric cancer
patients. Semin Pediatr Neurol. 2012;19:9–17.
6. Kerckhove N, Collin A, Condé S, Chaleteix C, Pezet D, Balayssac D. Long-term effects, pathophysiological mechanisms, and risk factors of chemotherapy-induced peripheral neuropathies: a comprehensive literature review. Front Pharmacol. 2017;24:86.
7. Gilchrist LS, Tanner LR, Ness KK. Short-term recovery of chemotherapy-induced peripheral neuropathy after treatment for pediatric non-CNS cancer
. Pediatr Blood Cancer
8. Lavoie Smith EM, Li L, Chiang C, et al Patterns and severity of vincristine-induced peripheral neuropathy in children with acute lymphoblastic leukemia. J Peripher Nerv Syst. 2015;20:37–46.
9. Ness KK, Hudson MM, Pui CH, et al Neuromuscular impairments in adult survivors of childhood acute lymphoblastic leukemia: associations with physical performance and chemotherapy doses. Cancer
10. Ramchandren S, Leonard M, Mody RJ, et al Peripheral neuropathy in survivors of childhood acute lymphoblastic leukemia. J Peripher Nerv Syst. 2009;14:184–189.
11. Shumway-Cook A, Woolacott MH. Motor Control: Translating Research Into Clinical Practice. Philadelphia, PA: Lippincott Williams & Wilkins; 2007.
12. Gilchrist LS, Tanner L. The pediatric-modified Total Neuropathy Score: a reliable and valid measure of chemotherapy-induced peripheral neuropathy in children with non-CNS cancers. Support Care Cancer
13. Tay CG, Lee VW, Ong LC, et al Vincristine-induced peripheral neuropathy in survivors of childhood acute lymphoblastic leukaemia. Pediatr Blood Cancer
14. Jain P, Gulati S, Seth R, et al Vincristine-induced neuropathy in childhood ALL (acute lymphoblastic leukemia) survivors: prevalence and electrophysiological characteristics. J Child Neurol. 2014;29:932–937.
15. De Luca CR, McCarthy M, Galvin J, et al Gross and fine motor skills in children treated for acute lymphoblastic leukaemia. Dev Neurorehabil. 2013;16:180–187.
16. Reinders-Messelink H, Schoemaker M, Snijders T, et al Motor performance of children during treatment for acute lymphoblastic leukemia. Med Pediatr Oncol. 1999;33:545–550.
17. Wright MJ, Galea V, Barr RD. Proficiency of balance in children and youth who have had acute lymphoblastic leukemia. Phys Ther. 2005;85:782–790.
18. Varedi M, McKenna R, Lamberg EM. Balance in children with acute lymphoblastic leukemia. Pediatr Int. 2017;59:293–302.
19. Hoffman MC, Mulrooney DA, Steinberger J, et al Deficits in physical function among young childhood cancer
survivors. J Clin Oncol. 2013;21:2799–2805.
20. Bruininks RH, Bruininks BD. BOT-2: Bruininks-Oseretsky Test of Motor Proficiency. 2nd ed. Minneapolis, MN: Pearson Assessments; 2005.
21. Wuang YP, Su CY. Reliability and responsiveness of the Bruininks-Oseretsky test of motor proficiency—second edition in children with intellectual disability. Res Dev Disabil. 2009;30:847–855.
22. Ness KK, DeLany JP, Kaste SC, et al Energy balance and fitness in adult survivors of childhood acute lymphoblastic leukemia. Blood. 2015;125:3411–3419.
23. Beulertz J, Bloch W, Prokop A, et al Limitations in ankle dorsiflexion range of motion, gait, and walking efficiency in childhood cancer
24. Hartman A, van den Bos C, Stijnen T, Pieters R. Decrease in peripheral muscle strength and ankle dorsiflexion as long-term side effects of treatment for childhood cancer
. Pediatr Blood Cancer
25. Ness KK, Krull KR, Jones KE, et al Physiologic frailty as a sign of accelerated aging among adult survivors of childhood cancer
: a report from the St Jude Lifetime cohort study. J Clin Oncol. 2013;31:4496–4503.