PR AMR falls and cancer treatment full article

fsc

lVto

dcsniv

P

M

1

ORIGINAL ARTICLE

Identifying Factors Associated With Falls in PostmenopausalBreast Cancer Survivors: A Multi-Disciplinary ApproachKerri M. Winters-Stone, PhD, Britta Torgrimson, PhD, Fay Horak, PhD, PT, Alvin Eisner, PhD,
Lillian Nail, PhD, RN, Michael C. Leo, PhD, Steve Chui, MD, Shiuh-Wen Luoh, MD
i

B

hacccw

ABSTRACT. Winters-Stone KM, Torgrimson B, Horak F,Eisner A, Nail L, Leo MC, Chui S, Luoh S-W. Identifyingfactors associated with falls in postmenopausal breast cancersurvivors: a multi-disciplinary approach. Arch Phys Med Re-habil 2011;xx:xxx.

Objective: To identify neuromuscular, balance, and visionactors that contribute to falls in recently treated breast cancerurvivors (BCS) and explore links between fall risk factors andancer treatment.

Design: Case-control plus prospective observation.Setting: Comprehensive cancer center.Participants: BCS (N�59; mean age, 58y) within 2 years of

chemotherapy completion and/or on adjuvant endocrine therapy.Interventions: Not applicable.Main Outcome Measures: Objective measures of postural

control, vision, and neuromuscular function included: (1) asensory organization test (SOT), (2) a visual assessment bat-tery, (3) muscle mass by dual energy x-ray absorptiometry, and(4) neuromuscular function with strength by repetition maxi-mum, power by timed stair climb, and gait speed by 4m walk.Falls were self-reported for the past year (retrospective) andmonthly for 6 months (prospective).

Results: Fifty eight percent of BCS reported falls in the pastyear. BCS with a history of falls had lower SOT scores with avestibular deficit pattern in postural control (P�.01) and tookonger to read letters on the contrast sensitivity chart (P�.05).estibular score on the SOT mediated the relationship between

reatment and falls among BCS who received chemotherapynly, but not adjuvant endocrine therapy.Conclusions: Results of this project suggest that balance

isturbances of vestibular origin and delays in detecting lowontrast visual stimuli are associated with falls in BCS. Futuretudies that track falls and fall risk factors in BCS from diag-osis through treatment are warranted, as are studies that candentify treatment-related vestibular dysfunction and alteredisual processing.Key Words: Drug therapy; Muscle strength; Neoplasms;

ostural balance; Rehabilitation.© 2011 by the American Congress of Rehabilitationedicine

From the Schools of Nursing (Winters-Stone, Torgrimson, Leo, Nail), and Medi-cine (Chui, Eisner, Horak, Luoh), Oregon Health and Science University; and thePortland Veterans Affairs Medical Center (Luoh), Portland, OR.

This publication was made possible with support from the Oregon Clinical andTranslational Research Institute (OCTRI), grant number UL1 RR024140 from theNational Center for Research Resources (NCRR), a component of the NationalInstitutes of Health (NIH), and NIH Roadmap for Medical Research.

No commercial party having a direct financial interest in the results of the researchsupporting this article has or will confer a benefit on the authors or on any organi-zation with which the authors are associated.

Reprint requests to Kerri M. Winters-Stone, PhD, Associate Professor, OregonHealth & Science University, 3455 SW US Veteran’s Hospital Rd, Mailcode: SN-ORD, Portland, OR 97239, e-mail: [email protected].

0003-9993/11/xxx-00851$36.00/0doi:10.1016/j.apmr.2010.10.039

FRACTURES ARE A PARTICULAR concern for BCS1,2

because of the combined effect of chemotherapy and ad-juvant endocrine therapy on fracture risk factors, includingbone mass2 and fall risk.3 Even though falls may play anmportant role in fracture etiology,4 there are few studies of

falls in BCS. We reported that BCS with chemotherapy-induced amenorrhea fall more than their age-matched, cancer-free peers.5 Chen et al3 reported an increased number of falls in

CS using data from the Women’s Health Initiative cohort.In older adults, leading risk factors for falls include a prior

istory of falls, muscle weakness, gait and/or balance impairment,nd poor vision.6-8 Cancer treatment may increase fall risk be-ause treatment can worsen traditional risk factors, potentiate lessommon risk factors, or contribute to falls in new ways. Breastancer chemotherapy may cause peripheral neuropathy9 andeakness.10 Adjuvant endocrine therapy has unknown effects on

neuromuscular function and may alter vision in several ways.11-15

Despite study limitations that included examination of few fallrisk factors and use of simple balance and strength assessments,our prior work suggests that both balance and strength are worsein estrogen-deficient BCS compared to controls and that falls aremore common in BCS with muscle weakness.5

The purpose of this study was to compare BCS with andwithout a recent history of falls on a comprehensive battery ofobjective tests of fall risk factors. The objectives of this studywere to: (1) identify balance, neuromuscular, and visual factorsassociated with past history of falls in BCS, (2) determine iffactors that discriminate past fallers from nonfallers can predictfuture falls, and (3) examine whether balance, neuromuscular,and visual factors are potential mediators of the relationshipbetween the type of cancer treatment and falls.

METHODS

Sample and SettingWomen were recruited through state and hospital cancer

registries, clinician referral, and community events. To bestevaluate persistent treatment effects rather than acute sideeffects of cancer chemotherapy that might be linked to falls,eligibility criteria included a diagnosis of nonmetastatic breastcancer (stages 0–IIIa), completion of chemotherapy within thepast 2 years with or without adjuvant hormone therapy orwithin 2 years of surgery and presently taking a SERM or AIthat was initiated at least more than 6 months earlier, post-menopausal, and less than 70 years of age (to separate treat-ment effects from aging). Women were excluded if they hadcognitive difficulties or medical conditions that, in the opinion

List of Abbreviations

AI aromatase inhibitorBCS breast cancer survivorsSERM selective estrogen receptor modulator
SOT sensory organization test
Arch Phys Med Rehabil Vol xx, Month 2011

mailto:[email protected]

dt

a

es3

d

sW

rtoDS

b

2 FALLS IN BREAST CANCER SURVIVORS, Winters-Stone

of the study investigator (K.W.S.), affected testing protocols. Aconvenience sample size of 60 participants was targeted for thispreliminary study, balancing available resources, while retain-ing power for regression models. The institutional reviewboard approved the study.

ProceduresWomen who responded to invitation letters from the can-

cer registry, advertisements, or oncologist referral were con-tacted by telephone and screened for eligibility. Eligiblewomen who agreed to participate provided written informedconsent, completed questionnaires and underwent balance,neuromuscular, and vision testing. Tests were performed bythe same technician except for the SOT, which was admin-istered by 1 of 2 trained physical therapists. All tests fol-lowed standard operating procedures that included moderateverbal encouragement. Participants also completed mailedor e-mailed monthly fall reports for the next 6 months. Latereports (�2 weeks) were followed up by phone and infor-mation was obtained verbally.

MeasuresDemographic and clinical characteristics. Demographic

and clinical characteristics including breast cancer stage, treat-ment type, diagnosis, and treatment dates were obtained byself-report.

Falls. Falls were assessed retrospectively through self-re-port of falls in the previous year and prospectively for 6 monthsby monthly reports. A fall was defined as unintentionallycoming to rest on the ground or at some other lower level, notas a result of a major intrinsic event or overwhelming hazard.16

A past fall was only counted if it occurred within the past yearand after cancer diagnosis. To reduce reporting bias, we used astandard definition of a fall, gave identical instructions fortracking falls to all participants regardless of fall history, andconfirmed affirmative fall reports by telephone.

Balance problems. Balance problems consistent with neu-ropathy, vestibular loss, and central nervous system disordersthat contribute to fall risk17 were assessed by computerizedynamic posturography (Neurocom Clinical Research Sys-ema). We conducted the SOT that systematically removes or

alters useful visual and somatosensory input across 6 sensory

Fig 1. Comparison of equilibrium scores by condition on the base-line SOT between BCS with a positive or negative fall history.

mFigure illustrating the sensory system alterations are placed beloweach condition number.


conditions (fig 1) and provides equilibrium scores for eachcondition based on center of gravity sway. Sensory analysisratios are calculated to identify the degree to which a personcan use information from the somatosensory, visual, and ves-tibular sensory systems to maintain balance, by comparing con-ditions 2 to 1, 4 to 1, and 5 to 1, respectively. The preference ratiodetermines reliance on visual information to maintain balanceeven when the visual information is incorrect from sway-refer-encing the visual surround (conditions 3 vs 6 and 2 vs 5).

Gait speed. Gait speed (m/s) was evaluated by timing theparticipant as she walked in a straight line for 4m at bothnormal and fast walking speeds on an electronic gait mat(GAITrite electronic walkway systemb). Slow gait speed isssociated with increased fall risk.6

Neuromuscular function. Neuromuscular function wasvaluated by the following tests: (1) maximal lower-bodytrength by a 1-repetition maximum leg press (kg) (Vectra VX8 multi station weight machinec). using standard protocols18;

(2) functional lower-body strength by length of time to rise 5times from a chair (s)19; and (3) functional stair climb ability,a task that relies on muscular power, by time to ascend andtime to descend 7 stairs (1.2m) using standard protocols20 andescribed as stair climb speed (m/s) and power (kg · m/s).Muscle mass. Muscle mass of the whole body was mea-

ured via dual energy x-ray absorptiometry (Hologic Discoveryi densitometerd). Participants lay supine on the table and

were positioned according to manufacturer guidelines for re-gional body composition analysis.d Body composition wasexpressed as percent whole-body lean mass and leg lean mass(kg). In-house coefficients of variation for lean and fat mass inour laboratory are less than 2%.21

Vision characteristics. Vision characteristics were as-sessed with a visual function battery consisting of the followingtests: (1) visual acuity using autorefraction (Humphrey Autore-fractor/Keratometer Model 599e). (2) spatial contrast sensitiv-ity (both eyes together) using the Pelli-Robson chart at aluminance level of 40cd/m2.22,f To assess task difficulty, weecorded Time to Hesitation as the time from the start of theest at highest contrast until subject first hesitated at 5 secondsr longer between successive letters at lower contrasts. (3)epth perception for near vision was assessed using the RandotO-002 Graded Circle Stereoacuity Testg; (4) visual field sen-

sitivity was measured using a white-on-white central 30°screening test (Humphrey Field Analyzer II model 750ie).

Statistical AnalysisDescriptive data for continuous variables are presented as

means and SDs. Categorical data are presented as percentage ofthe sample. Independent samples t tests or chi-square wereused for comparisons on balance, neuromuscular, vision, andtreatment characteristics between BCS with and without ahistory of falls. To explore predictors of falls over and above ahistory of past falls, hierarchical logistic multiple regressionwas used to examine treatment, balance, neuromuscular, andvision variables that predicted fall status in the 6-month fol-low-up period with variables that differed between groups at Pless than .15 considered for inclusion in the logistic regressionmodel as independent variables. Treatment was dummy codedto examine the independent contributions of chemotherapy andadjuvant endocrine therapy to fall prediction. To examinewhether balance, neuromuscular, and visual characteristics me-diate the relationship between treatment and falls, we em-ployed the product coefficients approach to testing media-tion.23-25 Only the variables that were significantly differentetween fallers and nonfallers at baseline were explored as
ediators.

F

w6obrf(cwmwfid

D

sf(sr

hst

P

awhocs(

NPi*

3FALLS IN BREAST CANCER SURVIVORS, Winters-Stone

RESULTS

allsOf the 143 women who expressed interest in the study, 59

omen enrolled and all completed baseline testing and-month fall reports. On average, participants were 58 yearsld and overweight (table 1). Most women had stage I or IIreast cancer, were treated with chemotherapy and were cur-ently taking a SERM or an AI. At enrollment, 58% reported aall in the previous year, and half experienced more than 1 falltable 2). Less than 10% of the sample had a fracture after theirancer diagnosis, but most reported fractures were associatedith a fall. Fifty-three percent of our sample experienced 1 orore falls during the 6-month follow-up period. One third ofomen with a previous fall history also fell in the 6-month

ollow-up period. In this study, taking both observation periodsnto account, 76% of BCS reported falling after their canceriagnosis within the 18-month time frame.

ifferences Between Past Fallers and NonfallersPast fallers and nonfallers were of similar age and clinical

tatus regarding breast cancer history (tables 3 and 4). Pastallers had lower scores on condition 5 of the SOT (P�.01)see fig 1), which contributed to differences on the vestibularcore (P�.01) (see table 4). Subjects who took longer time toead letters on the contrast sensitivity chart were more likely to

Table 1: Health History of Breast Cancer Survivors (N�59)

TreatmentMean � SD or% of Sample

Age (y) 58.5�9.7BMI (kg/m2) 28.3�7.2Months since diagnosis 20.9�6.5Stage 0 5Stage I 29Stage II 39Stage III 19Chemotherapy only (n�17) 29Current adjuvant endocrine therapy (n�19)

SERM use only (n�7) 12Current AI use only (n�12) 20

Chemotherapy and current SERM or AIuse (n�23) 39

Abbreviation: BMI, body mass index.

Table 2: Proportion of Falls and Fractures Reported byBreast Cancer Survivors (N�59) in the Past Year* and

Over the 6-Month Observation Period

Fall or Fracture Incidents % of Sample

At least 1 fall in past year 58Multiple falls in past year 31Recent fracture 8Recent fracture related to fall 5At least 1 fall occurring over 6-month observation

period 53Multiple falls occurring over 6-month observation

period 14At least 1 fall in past year or over 6-month

observation period 76

*Fall or fracture must have occurred within the year prior to enroll-ment and after cancer diagnosis.

ave a history of falls (P�.05) (table 5). There were no otherignificant differences on the SOT, the muscle function tests, orhe visual tests by fall history.

redictors of Future FallsAfter adjusting for fall history, the following variables were

dded in the regression model: vestibular score, vision score,hole body percent lean mass, stair descent speed, time toesitation, and treatment type. Though stair ascent speed metur criterion for entry in the model (P�.15), it was not in-luded in the regression because of its large correlation withtair descent speed and concern for multicollinearity. Time tovisual) hesitation and stair descent speed introduced instability

Table 3: Comparison of Participants With Positive and NegativeFall History* on Demographic and Clinical Information Collected

at Enrollment

Characteristic

Positive FallHistory(n�34)

Negative FallHistory(n�25) P

Age (y) 59.2�7.1 57.4�12.4 0.52Time since diagnosis (mo) 22.3�5.8 19.0�7.1 0.41Breast cancer stage 0.19

Stage 0 9 0 N/AStage I 27 37 N/AStage II 49 32 N/AStage III 15 31 N/A

TreatmentChemotherapy only 29 28 0.46Current SERM use only 15 4 0.28Current AI use only 15 24 0.16Chemotherapy and

SERM or AI use 41 44 0.72

OTE. Values are mean � SD, % of group, or as otherwise noted.ercentages reflect the proportion of the sample possessing a clin-

cal characteristic within each fall group.Number of falls reported over 12 months prior to baseline testing.

Table 4: Comparison of Participants With Positive andNegative Fall History* on Baseline Neuromuscular and

Balance Characteristics

Characteristic



BMI (kg/m2) 29.3�8.3 26.9�5.1 0.22Whole body % lean mass 60.7�6.8 63.4�6.2 0.12Leg lean mass (kg) 15.0�3.3 14.4�2.4 0.46Maximal leg press (kg) 86.0�30.4 78.6�19.5 0.30Chair stand time (sec to

complete 5 stands) 12.6�3.0 11.8�3.3 0.31Stair climb power

(kg · m/s) 28.3�6.7 28.2�7.0 0.96Stair ascent speed (m/s) 0.25�0.04 0.27�0.05 0.09Stair descent speed (m/s) 0.28�0.07* 0.31�0.05 0.07Best fast walk (m/s) 0.43�0.07 0.48�0.07 0.21Average usual walk (m/s) 0.30�0.05 0.33�0.05 0.21SOT-somatosensory 94.6�7.2 95.6�3.5 0.52SOT-visual 77.6�19.1 84.1�8.5 0.12SOT-vestibular 41.4�25.0 57.3�18.7 0.01SOT-preference 96.6�23.6 98.6�11.5 0.70

NOTE. Values are mean � SD or as otherwise noted.
Abbreviation: BMI, body mass index.*Number of falls reported over 12 months prior to baseline testing.

t1ctm

f

smr

awseDtd

ilhsnWamcvmar

tnh


in the estimates of the SEs so the model was rerun without thesevariables. The final logistic regression model was not significant(�2

6� 3.85, P�.05) (table 6).

Mediators of the Relationship Between Treatmentand Falls

Vestibular score was the only variable that met criteria formediator models. Separate models were run to examinewhether vestibular score mediated the relationship betweenpast or future falls and a particular treatment category (adjuvantendocrine therapy, chemotherapy, or adjuvant endocrine ther-apy and chemotherapy). Treatment type had an indirect rela-tionship with past falls due to balance problems when relyingon the vestibular system (table 7). The chemotherapy onlygroup was 1.96 times more likely to have a history of falls thanthe group that received chemotherapy and endocrine therapyvia the effect of treatment on vestibular score (95% bias-corrected and accelerated confidence interval for the oddsratio�1.05–6.38). Vestibular score did not mediate the rela-tionship between treatment type and future falls.

DISCUSSIONOur study is the first to consider how breast cancer treatment

may increase fall risk by using a comprehensive set of objec-tive measures of fall risk and by exploring mediators of thetreatment-falls relationship. Our findings suggest that recentlytreated postmenopausal BCS have higher rates of falling com-pared with population averages for community-dwelling olderadults. Balance disturbances may explain how treatment couldincrease falls in BCS. Difficulties responding quickly to lowspatial contrasts may also contribute to falls in this population,in which vision disturbances associated with adjuvant endo-crine therapy have been reported.11-15 By aggregating fall his-ory data and prospective fall data, 76% of BCS fell within an8-month period. Alterations in vestibular inputs to balanceontrol may explain the relationship between breast cancerreatment and falls observed in women whose systemic treat-ent was cytotoxic chemotherapy alone.Among our sample of postmenopausal BCS, the incidence of

Table 5: Comparison of Participants With Positive and NegativeFall History* on Baseline Vision Characteristics

Characteristic



Visual acuity – right eye† 0.9�0.2 0.9�0.3 0.63Visual acuity – left eye† 0.6�0.7 0.6�0.5 0.89Contrast sensitivity‡ 1.8�0.2 1.8�0.2 0.73Contrast sensitivity time to

hesitation (s)§ 29.0�10.5 24.6�6.1 0.04Contrast sensitivity at

hesitation§ 1.7�0.1 1.7�0.1 0.99Stereovision (arcsecond)� 88.5�169.0 173.0�314.7 0.63Visual field sensitivity (no.

of missed points)¶ 0.8�2.6 3.8�8.5 0.10

NOTE. Values are mean � SD or as otherwise noted.*Number of falls reported over 12 months prior to baseline testing.†Scores � 1 indicate poorer than 20/20 acuity.‡Higher scores indicate better contrast sensitivity.§Higher scores indicate more delay in detecting letters at low con-rast; sample sizes for contrast sensitivity were positive fall history:�31 and negative fall history: n�23 and for contrast sensitivityesitation time and line were n�29 and n�22 due to missing data.

�Lower scores (arcsecond) indicate better stereovision.¶Higher scores imply poorer coverage of the visual field.

alls is nearly double the 25% to 30% annual fall rate reported for


community-dwelling older adults over 65 years of age.26 In ourtudy, 58% of BCS experienced a fall in the year prior to enroll-ent and nearly half (47%) fell within 6-months thereafter. Our

esults agree with those reported by Chen et al,3 that BCS have a15% higher risk of falls than non-BCS. In contrast, studies in oldercancer survivors report similar fall histories between older adultsreceiving treatment for cancer and cancer-free older adults.27,28

Age and treatment differences may explain the discrepanciesbetween falls data in BCS studies versus studies of older adultswith mixed cancer diagnoses. At least in BCS, higher fall riskshould be considered as a potential consequence of treatment. Inturn, falls may lead to injury, activity limitations due to fear offalling,29 and eventually contribute to functional decline.30

We measured a comprehensive set of neuromuscular and bal-ance characteristics known to be associated with falls, but onlybalance discriminated fallers from nonfallers. BCS who fell hadsignificantly lower scores on condition 5 of the SOT, indicating avestibular-related deficit in balance control. Using similar meth-ods, Wampler et al31 reported lower scores on 5 of 6 SOT tests,and particularly lower vestibular scores among BCS treated withtaxane therapy compared with healthy, age-matched controls.SOT scores from our BCS group are similar to the BCS sample ofthe Wampler31 study. However, in our study, the vestibular scoresmong the fallers are even lower and sway in other conditionsas not abnormal. The similar SOT results for vestibular

cores reported by both studies may reflect a consistent sideffect of breast cancer treatment, particularly chemotherapy.oxobubicin (adriamycin), used in most combination chemo-

herapy for breast cancer, is an aminoglycoside. This class ofrugs is associated with vestibular ototoxicity.32 Ototoxic side

effects from chemotherapy drugs for breast cancer have re-ceived very little attention in the literature. Of the 2 studiesincluding BCS, aminoglycosides were linked to ototoxicity inBCS only,33 while ototoxicity from taxanes was not observedn breast plus ovarian cancer survivors.7 Ototoxicity results inoss of hair cells in the vestibular system even before noticeableair cell loss in the cochlea for hearing and consistently re-ults in abnormal scores in SOT conditions 5 and 6, despiteormal sway in conditions 1 through 4.34 This study and theampler31 study suggest that breast cancer treatment may

dversely affect the vestibular inputs to balance control and thisay lead to falls. However, because past, but not future falls,

ould be predicted by balance problems, it is possible thatestibular or other deficits impairing balance function wereost severe during and immediately following chemotherapy

nd lessen over time or are compensated for by increasedeliance on other sensory systems to maintain balance.35

Table 6: Logistic Regression Results for Prediction of Faller StatusOver 6-Month Observation Period

Independent Variable Odds Ratio (95% CI) P

History of falls in previous year 0.86 (0.27–2.69) 0.79Whole body percent lean mass 1.00 (0.92–1.01) 0.94SOT-vision 0.97 (0.93–1.01) 0.17SOT-vestibular 0.99 (0.97–1.03) 0.77Adjuvant endocrine therapy only* 1.42 (0.41–4.90) 0.58Chemotherapy only† 1.32 (0.31–5.64) 0.71

NOTE. Independent variables were assessed at baseline.Abbreviation: CI, confidence interval.*Comparison of participants who only had adjuvant endocrine ther-apy to participants who had chemotherapy and adjuvant endocrinetherapy.†
Comparison of participants who had chemotherapy only compared toparticipants who had chemotherapy and adjuvant endocrine therapy.

mes


Wampler et al also reported that BCS treated with taxane-based chemotherapy had worse visual acuity at low- but nothigh-contrast, as compared with healthy control subjects. In ourstudy, the average contrast sensitivity level itself was notsignificantly different between fallers and nonfallers, whichcould be because the Pelli-Robson test assesses contrast sen-sitivity far above the visual acuity limit. In addition, we foundthe relation of falls to contrast sensitivity to depend on the timetaken by subjects to respond correctly. In the absence of timeconstraints, subjects may view the low-contrast letters for up-wards of 10 seconds before identifying the correct letters. Adecision period this long would be of little help to a personnavigating real-world environments where falls occur. Ourresults may help explain discrepancies in the literature on therelation of contrast sensitivity to falls.36-38

In the mediator models, chemotherapy treatment had anindirect effect on falls through its effect on vestibular balancescores when contrasted with the group that received chemo-therapy and adjuvant endocrine therapy where no relationshipwith vestibular balance scores and falls existed. However, therewas no indirect effect found for adjuvant endocrine therapyonly and falls via vestibular scores when compared to thosewho received chemotherapy and adjuvant endocrine therapy.The differences between treatment groups may be confoundedby different health and disease status of the patient that dictatedtype of treatment. Women who do not receive adjuvant endo-crine therapy for breast cancer are more likely to have hormonereceptor negative disease that is not responsive to SERMs orAIs.39 These BCS may be more likely to receive high-dosechemotherapy40 and/or anthracycline-based chemotherapy41

thus, may be more vulnerable to side effects of this type ofchemotherapy treatment. Given the variation in dose and typeof chemotherapy and the impact of these variations on thenature and magnitude of treatment-related side effects, it willbe important for future studies to obtain complete treatmentinformation on participants and/or focus on specific chemo-therapeutic agents.

Study LimitationsOur study has limitations. The analyses using retrospective

Table 7: Mediation Analysis to Test for the Indirect Effects

Pathway

Test of indirect path from adjuvant endocrine therapy only† as theAdjuvant endocrine therapy only to VEST scoreDirect effects of VEST score on fallsTotal effect of adjuvant endocrine therapy only on fallsDirect effect of adjuvant endocrine therapy only on fallsPartial effect of chemotherapy only on falls‡

Indirect effects of adjuvant endocrine therapy only on falls throuTest of indirect path from chemotherapy only‡ as the independen

Chemotherapy only to VEST scoreDirect effects of VEST score on fallsTotal effect of chemotherapy only on fallsDirect effect of chemotherapy only on fallsPartial effect of adjuvant endocrine therapy only on falls†

Indirect effects of chemotherapy only on falls through VEST sco

Abbreviations: CI, confidence interval; N/A, not applicable; VEST, ve*Odds ratios reported only for logistic regression analyses.†Comparison of participants who only had adjuvant endocrine therapy‡Comparison of participants who had chemotherapy only compared§Normal theory tests are not available for models with binary outco�Confidence intervals are bias corrected and accelerated.

falls data are cross-sectional and cannot establish cause and

effect; in fact, we only counted falls that occurred in the pastyear and after cancer diagnosis to retain the plausible link tocancer treatment. Due to challenges inherent to abstractingmedical records from diverse hospital systems among ourcommunity sample, we were unable to obtain full informationon chemotherapy drugs and doses received by participants. Asa result, we cannot link specific chemotherapy agents to falls.We did not include measures of peripheral neuropathy, a com-mon side effect of chemotherapy associated with falls; how-ever, the SOT somatosensory scores were similar betweenfallers and nonfallers (see fig 1), suggesting that peripheralneuropathy may not underlie balance problems in this sample.Our study may have attracted women concerned about fallingand results may not generalize to the broader population ofBCS. Our sample size was relatively small but appropriate forpreliminary work and intended to inform future studies.

CONCLUSIONS

Falls in BCS are understudied and deserve more attention,particularly in light of the increase in fractures after breastcancer treatment and the relationship of falls to fractures. Ourfindings add to growing evidence that fall risk is increased inBCS and that vestibular function may underpin associationsbetween breast cancer treatment and falls. Our results alsosuggest that standard contrast sensitivity tests may not capturedelays in contrast detection that might contribute to falls.Collectively, knowledge from this study will inform futurestudies of falls in cancer survivors as well as contribute to thedevelopment of specific interventions that target BCS at risk offalling. To date, only strength training has been examined foreffects on falls in BCS.42 This study did not specifically targetBCS who fell and exercise did not reduce falls. Given ourfindings suggesting that strength is not a predictor of falls inBCS, strength training may be less effective in BCS than inolder adults where strength is a predictor of falls. Future studiesmay consider multifactorial interventions for reducing falls thatinclude vestibular rehabilitation techniques and environmental

atment Type on Falls Via Vestibular Scores From the SOT

b SE P Odds Ratio*95% CI

Odds Ratio

pendent variable–6.43 6.84 0.35 N/A N/A–0.04 0.01 0.02 0.96 0.94–0.980.47 0.61 0.45 1.60 0.48–5.290.31 0.65 0.63 1.36 0.38–4.87

–0.23 0.78 0.77 0.79 0.17–3.66EST score 0.22 0.30 N/A§ 1.25 0.76–2.64�

able–19.41 7.72 0.02 N/A N/A–0.04 0.01 0.02 0.96 0.94–0.980.40 0.69 0.57 1.49 0.39–5.77

–0.23 0.78 0.77 0.79 0.17–3.660.31 0.65 0.63 1.36 0.38–4.870.68 0.49 N/A§ 1.96 1.05–6.38�

lar.

articipants who had chemotherapy and adjuvant endocrine therapy.rticipants who had chemotherapy and adjuvant endocrine therapy.

.

of Tre

inde

gh Vt vari

re

stibu

to pto pa

modifications to enhance contrast.


2

2

2

2

2

2

2

2

3

3

3

3

3

3

3

3

3

3

4

4

4


References1. Kanis JA, McCloskey EV, Powles T, Paterson AH, Ashley S,

Spector T. A high incidence of vertebral fracture in women withbreast cancer. Br J Cancer 1999;79:1179-81.

2. Howell A, Cuzick J, Baum M, et al. Results of the ATAC(Arimidex, Tamoxifen, Alone or in Combination) trial after com-pletion of 5 years’ adjuvant treatment for breast cancer. Lancet2005;365:60-2.

3. Chen Z, Maricic M, Bassford TL, et al. Fracture risk among breastcancer survivors: results from the Women’s Health InitiativeObservational Study. Arch Int Med 2005;165:552-8.

4. Frost HM. Should fracture risk-of-fracture analyses include an-other major risk factor? The case for falls. J Clin Densitom2001;4:381-3.

5. Winters-Stone KM, Nail L, Bennett JA, Schwartz A. Bone healthand falls: fracture risk in breast cancer survivors with chemother-apy-induced amenorrhea. Oncol Nurs Forum 2009;36:315-25.

6. Guideline for the prevention of falls in older persons. AmericanGeriatrics Society, British Geriatrics Society, and American Acad-emy of Orthopaedic Surgeons Panel on Falls Prevention. J AmGeriatr Soc 2001;49:664-72.

7. Sarafraz M, Ahmadi K. Paraclinical evaluation of side-effects oftaxanes on auditory system. Acta Otorhinolaryngol Ital 2008;28:239-42.

8. Dhital A, Pey T, Stanford MR. Visual loss and falls: a review. Eye(Lond) 2010;24:1437-46.

9. Kuroi K, Shimozuma K. Neurotoxicity of taxanes: symptoms andquality of life assessment. Breast Cancer 2004;11:92-9.

10. Mold JW, Vesely SK, Keyl BA, Schenk JB, Roberts M. The prev-alence, predictors, and consequences of peripheral sensory neuropa-thy in older patients. J Am Board Fam Pract 2004;17:309-18.

11. Eisner A, Austin DF, Samples JR. Short wavelength automatedperimetry and tamoxifen use. Br J Ophthalmol 2004;88:125-30.

12. Epstein RJ. Visual impairment in myopic patients with breastcancer receiving adjuvant therapy with aromatase inhibitors. ClinBreast Cancer 2009;9:184-6.

13. Eisner A, Incognito LJ. The color appearance of stimuli detectedvia short-wavelength-sensitive cones for breast cancer survivorsusing tamoxifen. Vision Res 2006;46:1816-22.

14. Eisner A, Toomey MD, Falardeau J, Samples JR, Vetto JT. Differ-ential effects of tamoxifen and anastrozole on optic cup size in breastcancer survivors. Breast Cancer Res Treat 2007;106:161-70.

15. Coombes RC, Hall E, Gibson LJ, et al. A randomized trial ofexemestane after two to three years of tamoxifen therapy inpostmenopausal women with primary breast cancer. New EnglJ Med 2004;350:1081-92.

16. Tinetti M, Williams C. Falls, injuries due to falls, and the risk ofadmission to a nursing home. N Engl J Med 1997;337:1279-84.

17. Whitney SL, Marchetti GF, Schade AI. The relationship betweenfalls history and computerized dynamic posturography in personswith balance and vestibular disorders. Arch Phys Med Rehab2006;87:402-7.

18. American College of Sports Medicine. ACSM’s guildelines forexercise testing and prescription. 7th ed. Philadelphia: LippincottWilliams & Wilkins; 2006.

19. Guralnik J, Simonsick E, Ferrucci L, et al. A short physicalperformance battery assessing lower extremity function: associa-tion with self-reported disability and prediction of mortality andnursing home admission. J Gerontol 1994;49:M85-94.

20. Bean JF, Kiely DK, LaRose S, Alian J, Frontera WR. Is stairclimb power a clinically relevant measure of leg power impair-ments in at-risk older adults? Arch Phys Med Rehab 2007;88:604-9.

21. Winters KM, Snow CM. Detraining reverses positive effects ofexercise on the musculoskeletal system in premenopausal women.
J Bone Miner Res 2000;15:2495-503.

2. Rubin GS, West SK, Munoz B, Bandeen-Roche K, Zeger S,Schein O, Fried LP. A comprehensive assessment of visual im-pairment in a population of older Americans. The SEE Study.Salisbury Eye Evaluation Project. Invest Ophthalmol Vis Sci1997;38:557-68.

3. Hayes AF. Beyond Baron and Kenny: statistical mediation anal-ysis in the new millennium. Communication Monographs 2009;76:408-20.

4. MacKinnon DP, Lockwood CM, Williams J. Confidence limits forthe indirect effect: distribution of the product and resamplingmethods. Multivariate Behav Res 2004;39:99.

5. MacKinnon DP, Lockwood CM, Hoffman JM, West SG, SheetsV. A comparison of methods to test mediation and other interven-ing variable effects. Psych Methods 2002;7:83-104.

6. Hornbrook MC, Stevens VJ, Wingfield DJ, Hollis JF, GreenlickMR, Ory MG. Preventing falls among community-dwelling olderpersons: results from a randomized trial. Gerontologist 1994;34:16-23.

7. Overcash JA, Beckstead J. Predicting falls in older patients usingcomponents of a comprehensive geriatric assessment. Clin J On-col Nurs 2008;12:941-9.

8. Overcash J. Prediction of falls in older adults with cancer: apreliminary study. Oncol Nurs Forum 2007;34:341-6.

9. Salkeld G, Cameron I, Cumming R, et al. Quality of life related tofear of falling and hip fracture in older women: a time trade offstudy. BMJ 2000;320:341-6.

0. Deshpande N, Metter EJ, Lauretani F, Bandinelli S, Guralnik J,Ferrucci L. Activity restriction induced by fear of falling andobjective and subjective measures of physical function: a prospec-tive cohort study. J Am Geriatr Soc 2008;56:615-20.

1. Wampler MA, Topp KS, Miaskowski C, Byl NN, Rugo HS,Hamel K. Quantitative and clinical description of postural insta-bility in women with breast cancer treated with taxane chemo-therapy. Arch Phys Med Rehab 2007;88:1002-8.

2. Black FO, Pesznecker SC. Vestibular ototoxicity. Clinical consid-erations. Otolaryngol Clin North Am 1993;26:713-36.

3. Jenkins V, Beveridge H, Low R, Mitra S. Atypical hearing loss inwomen with breast cancer receiving adjuvant treatment. Breast2006;15:448-51.

4. Black FO. What can posturography tell us about vestibular func-tion? Ann N Y Acad Sci 2001;942:446-64.

5. Horak FB. Postural compensation for vestibular loss and implica-tions for rehabilitation. Restor Neurol Neurosci 2010;28:57-68.

6. de Boer MR, Pluijm SM, Lips P, et al. Different aspects of visualimpairment as risk factors for falls and fractures in older men andwomen. J Bone Min Res 2004;19:1539-47.

7. Lord SR. Visual risk factors for falls in older people. Age Ageing2006;(35 Suppl 2):ii42-ii45.

8. Freeman EE, Munoz B, Rubin G, West SK. Visual field lossincreases the risk of falls in older adults: the Salisbury eye eval-uation. Invest Ophthalmol Vis Sci 2007;48:4445-50.

9. Chen JQ, Russo J. ERalpha-negative and triple negative breastcancer: molecular features and potential therapeutic approaches.Biochim Biophys Acta 2009;1796:162-75.

0. Nieto Y, Shpall EJ. High-dose chemotherapy for high-risk pri-mary and metastatic breast cancer: is another look warranted?Curr Opin Oncol 2009;21:150-7.

1. Berry DA, Cirrincione C, Henderson IC, et al. Estrogen-receptorstatus and outcomes of modern chemotherapy for patients withnode-positive breast cancer. JAMA 2006;295:1658-67.

2. Twiss JJ, Waltman NL, Berg K, Ott CD, Gross GJ, Lindsey AM.An exercise intervention for breast cancer survivors with bone
loss. J Nurs Scholarsh 2009;41:20-7.

a

bcd

e

f

g


Suppliers. NeuroCom International Inc, 9570 SE Lawnfield Rd, Clackamas,

OR 97015.. CIR Systems Inc, 60 Garlor Dr, Havertown, PA 19083.. Vectra Fitness, 7901 S 190th St, Kent, WA 98032.
. Hologic Inc, 35 Crosby Dr, Bedford, MA 01730.
. Carl Zeiss Meditec Inc, 5160 Hacienda Dr, Dublin, CA 94568-7562.

. Lombart Instrument, 5358 Robin Hood Rd, Norfolk, Virginia23513.

. Stereo Optical Co, 8623 W Bryn Mawr Ave, Ste 502, Chicago, IL
60631.

Health & Medicine

PR AMR falls and cancer treatment full article