Pal Et Al-2010-CA- A Cancer Journal for Clinicians

Evaluating the Older Patient with Cancer:Understanding Frailty and the Geriatric Assessment

Sumanta Kumar Pal, MD1; Vani Katheria, BSc2; Arti Hurria, MD3

AbstractThe majority of cancer incidence and mortality occurs in individuals aged older than 65 years, and the number ofolder adults with cancer is projected to significantly increase secondary to the aging of the US population. As such,understanding the changes accompanying age in the context of the cancer patient is of critical importance. Age-related changes can impact tolerance of anticancer therapy and can shift the overall risk-benefit ratio of suchtreatment. A challenge in implementing evidence-based approaches in older adults is the under–representation ofthis group in oncology clinical trials. In addition, although older adults are particularly vulnerable to the side effectsof cancer therapy, few oncology studies to date have incorporated a measure of health status other than theEastern Cooperative Oncology Group or Karnofsky performance scales. Novel metrics such as frailty indices or thegeriatric assessment recognize heterogeneity among older adults, and may allow for risk-adapted approaches totherapy. It is increasingly recognized that several laboratory markers may predict morbidity and mortality in olderadults; these biologic variables may further aid in stratifying this group of patients based on risk. This reviewdescribes key studies from the geriatric literature that provide principles for assessing health status in the olderpatient, and ways that these principles can be applied to oncology care in an older population are proposed. CACancer J Clin 2010;60:120–132. ©2010 American Cancer Society, Inc.

IntroductionEstimates from the US Census Bureau predict a rapid rise in the number of individuals aged older than 65 years.1

Furthermore, the number of individuals over the age of 65 years and 85 years is projected to double to nearly 62million and 10 million, respectively, by the year 2030 and centenarians are anticipated to be the fastest-growinggroup within this demographic.2 Of particular concern in this population is the association between cancer andaging. It is estimated that 1,479,350 new cases of cancer were diagnosed in 2009, and 562,340 deaths occurred.3

Of these, it is projected that approximately 60% of cancer incidence and 70% of cancer-related mortality will occurin individuals aged older than 65 years, with even more pronounced figures reported for specific diseases.4-6

Several unique concerns arise in the older adult with cancer. With increasing age, physiologic reserve decreases;however, the pace of this decline varies with each individual.7 Similarly, variations in functional status, cognition,and comorbidity accompany increased age, and may affect life expectancy, risk of subsequent functional decline,hospitalization, and other morbidity.8-11 These age-related changes can influence tolerance to cancer therapy, aswell as the overall risk-benefit ratio of cancer treatment. The rate of these changes also varies between individuals.

Although older adults have been identified as being vulnerable to side effects from cancer therapy, few oncologystudies to date have specifically incorporated baseline metrics for measuring health conditions other thanfunctional status (Eastern Cooperative Oncology Group [ECOG] or Karnofsky performance status [PS]) to

1Assistant Professor, Division of Genitourinary Malignancies, Department of Medical Oncology and Experimental Therapeutics, City of Hope ComprehensiveCancer Center, Duarte, CA; 2Cancer Control and Population Sciences Program, City of Hope Comprehensive Cancer Center, Duarte, CA; 3Associate Professor,Department of Medical Oncology and Experimental Therapeutics, Cancer Control and Population Sciences Program, and Director, Cancer and Aging ResearchProgram, City of Hope Comprehensive Cancer Center, Duarte, CA.

Corresponding author: Arti Hurria, MD, Department of Medical Oncology and Experimental Therapeutics, Cancer Control and Population Sciences Program, and Director,Cancer and Aging Research Program, City of Hope Comprehensive Cancer Center, Duarte, CA 91010; [email protected]

DISCLOSURES: The authors reported no conflicts of interest.

�2010 American Cancer Society, Inc. doi:10.3322/caac.20059.

Available online at http://cajournal.org and http://cacancerjournal.org

Frailty and the Geriatric Assessment

120 CA: A Cancer Journal for Clinicians

identify individuals most at risk.Furthermore, older adults havebeen under-represented in oncol-ogy clinical trials, and only a fewstudies published to date have fo-cused on individuals who are con-sidered physically unable to re-ceive standard cancer therapy.12,13

Because the oncology literatureprovides little data regarding thesubject, general principles fromthe geriatric literature can be usedto guide the oncologist in identi-fying those patients who are frailand at high risk for functional de-cline, hospitalization, institution-alization, and mortality.

In this review, we describe keystudies from the geriatric litera-ture, and propose an applicationof these studies to care in anolder oncology population. We also discuss potentialbiologic markers of aging, which provide an insightinto clinical manifestations.

Understanding FrailtyUntil recently, a consistent definition of frailty remainedelusive. However, emerging data are providing a uni-form definition of frailty within the geriatric population.A position statement from the American Medical As-sociation defined the term “frailty” as characterizing“the group of patients that presents the most complexand challenging problems to the physician and allhealth care professionals,” because these are the indi-viduals who have a higher susceptibility to adverse out-comes, such as institutionalization or mortality.14-16

Clinical Manifestations of Frailty

Phenotype Derived from the CardiovascularHealth StudyData from the Cardiovascular Health Study (CHS)have been used to construct a widely cited definition offrailty.15 Based on a proposed “cycle of frailty” (Fig. 1),the investigators posited that 5 variables estimated a“phenotype of frailty”: 1) shrinking (weight loss), 2)weakness, 3) poor endurance, 4) slowness, and 5) low

physical activity. Measurable parameters for each ofthese variables were developed, and are described inTable 1. With data from 5317 patients assessed, ap-proximately 7% of the study population was consideredfrail (meeting greater than 3 frailty criteria), and 47%were considered to be prefrail (meeting 1–2 frailty cri-teria). There was an association noted between frailtyand 5 adverse outcomes: 1) hospitalization, 2) falls, 3)worsening activity of daily living (ADL) disability, 4)worsening mobility disability, and 5) death (P � .0001for all) (Fig. 2). Patients currently undergoing treatmentfor cancer were excluded from enrollment. Neverthe-less, the number of patients with cancer (not receivingactive treatment) did not differ significantly betweenthe frail, prefrail, and nonfrail groups (14%, 15%, and16%, respectively; P � .42). As such, the effects of acancer diagnosis on frailty are difficult to assess in thisstudy. However, distinct demographic characteristicsassociated with the “frail population” were identified,including older age, female gender, and African Amer-ican race.17

Women’s Health and Aging StudiesThe Women’s Health and Aging Studies (WHAS) Iand II were used to further validate the results of theCHS.18 In the WHAS studies, previously establishedcriteria were used to stratify women into 1 of 3 tertiles

FIGURE 1. Cycle of Frailty. VO2max indicates maximum oxygen consumption. Adapted with permission fromFried LP, Tangen CM, Walston J, et al. Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci MedSci. 2001;56:M146–M157.15

CA CANCER J CLIN 2010;60:120–132

121VOLUME 60 � NUMBER 2 � MARCH/APRIL 2010

of disability.19 WHAS I enrolled women aged 65 yearsand older who were in the most disabled tertile amongthe community-dwelling population (N�1002). Incontrast, WHAS II enrolled women ages 70 to 79 yearswho were in the 2 least disabled tertiles (N�436). Witha slight modification of the 5 criteria derived from theCHS (Table 2), there was agreement to within 7% inthe frequency of frailty in the 2 studies. Furthermore,the criteria predicted mortality in the WHAS experi-ence. The Three-City Study conducted in France,however, yielded different results.20 Using the samecriteria, a similar proportion of frail individuals (7%)was identified in a population of 6078 community-dwelling older men and women. Although frailty wasfound to be associated strongly with the need for assis-tance with ADLs and instrumental ADLs (IADLs)(Table 3), it failed to predict adverse outcomes such ashospitalization or mortality.

In an effort to simplify the CHS model, the Studyof Osteoporotic Fractures (SOF) index was exam-ined. The SOF index includes just 3 components(weight loss, inability to rise from a chair, and poorenergy). In separate analyses of 3130 older males and6701 older females, the SOF index demonstratedpredictive capabilities for morbidity and mortalitysimilar to a slightly modified CHS frailty index.21,22

Retornaz et al assessed the 5 CHS frailty markers(Table 1) in addition to mood and cognition in 50chemotherapy recipients aged 70 years and older.23

Although this analysis did not provide data regardingclinical outcome based on these measures, it didappear that the use of this wide panel of frailtymakers identified a distinct group of patients com-pared with the assessment of ADL and IADL de-pendency alone. In fact, 42% of patients included inthe study had no ADL or IADL disability, butnonetheless had more than 1 frailty marker.

Bylow et al described a theoretical framework forthe development of frailty in the setting of androgendeprivation therapy (ADT) for prostate cancer.24 Ex-isting data support an association between ADT andeach of the 5 CHS frailty markers. For example,unintentional lean weight loss (as a consequence ofsarcopenia) frequently occurs in older adults receivingADT.25 Sarcopenia consequently leads to increases inweakness and decreases in mobility, 2 separate com-ponents of the CHS frailty model. Finally, ADT hasbeen associated with both fatigue and lowered phys-ical activity, which are the remaining components ofthe model.26 Prospective studies are needed to vali-date this theoretical framework.

FIGURE 2. Incidence of Adverse Outcomes Associated With Frailty. The3-year outcomes denoted here were adapted from Table 6 in Fried LP, TangenCM, Walston J, et al. Frailty in older adults: evidence for a phenotype. J GerontolA Biol Sci Med Sci. 2001;56:M146–M157.15

TABLE 1. Frailty Criteria Proposed in the Cardiovascular Health Study15

FRAILTY CRITERIA CHARACTERISTIC

Shrinking ● Weight loss (unintentional) of �10 lbs in prior y OR● Weight loss at follow-up of �5% of body weight in prior y (by direct measurement of weight).

Weakness ● Grip strength in the lowest 20% at baseline, adjusted for gender and BMI.

Poor enduranceand energy

● Poor endurance and energy as indicated by self-report of exhaustion.● Self-reported exhaustion, identified by 2 questions from the CES-D scale, is associated withstage of exercise reached in graded exercise testing as an indicator of VO2 max and is predictiveof cardiovascular disease.

Slowness ● Slowest 20% of the population was defined at baseline based on time to walk 15 ft, adjustingfor gender and standing height.

Low physical levelof activity

● Weighted score of Kcal expended per wk was calculated at baseline, based on eachparticipant’s report.●Lowest quintile of physical activity was identified for each gender.

BMI, body mass index; CES–D, Center for Epidemiologic Studies Depression Scale; VO2 max, maximum oxygen consumption; Kcal, kilocalories.



Canadian Study of Health and Aging

Data from the Canadian Study of Health and Aging(CSHA) were used to validate a separate set of frailtycriteria derived from the Geriatric Status scale. Thistool utilizes both functional and cognitive criteria inthe definition of frailty (Table 4).16 The CSHAincluded 9008 adults aged 65 years and older, with abaseline assessment and 5-year follow-up for adverseoutcomes. In this model, the prevalence of frailty was

lower than that identified by the CHS; specifically,frailty was identified in 0.7%, 2%, and 4%, respec-tively, of patients ages 65 to 74 years, 75 to 84 years,

TABLE 3. ADLs and Instrumental ADLs125

ADLS IADLS

● Bathing● Dressing● Toileting● Transferring● Continence● Feeding

● Ability to use telephone● Shopping● Food preparation● Housekeeping● Laundry● Mode of transportation● Ability to take own medications● Ability to handle finances

ADLs indicates activities of daily living; IADLs, instrumental ADLs.

TABLE 2. Frailty Defined by the Women’s Health and Aging Study and the Cardiovascular Health Study18

FRAILTYCRITERIA CHS WHAS

Weight loss Lost �10 lbs unintentionally in last y ● �10% weight loss compared with weight at age 60 yOR● BMI at examination �18.5 kg/m2

Exhaustion Self-report of either:● Feeling everything I did was an effort in the last wk● Could not get going in the last wk

Self-report of any:● Low usual energy level (�3; range, 0-10)a

● Felt unusually tired in the last mo● Felt unusually weak in the past mo

Low energyexpenditure

MLTA questionnaire (short version)● Evaluating all 18 items● �270 Kcals per wk on activity scale

MLTA questionnaire (short version)● Evaluating 6 of 18 items:

- Walking- Strenuous household chores- Strenuous outdoor chores- Dancing- Bowling- Exercise

● �90 Kcals per wk on activity scale

Slowness Walking 15 ft (4.57 m):● Time �7 s for height �159 cm● Time �6 s for height �159 cm

Walking 4 m:● Walking speed (m/s) same as the CHS criteria

Weakness Grip strength (in kg) in dominant hand, measured bya Jamar handheld dynamometer:● �17 for BMI �23 kg/m2

● �17.3 for BMI of 23.1-26 kg/m2

● �18 for BMI of 26.1-29 kg/m2

● �21 for BMI �29 kg/m2

Grip strength:● Same as CHS criteria

Overall frailtystatus

Robust: met none of the criteriaIntermediate: met 1 or 2 criteriaFrail: met �3 criteria

CHS indicates Cardiovascular Health Study; WHAS, Women’s Health and Aging Study; BMI, body mass index; MLTA, Minnesota Leisure Time Activity; Kcals,kilocalories.aRated on a scale of 0-10, in which 0 indicates no energy and 10 indicates the most energy that you have ever had.

TABLE 4. Frailty Markers Proposed in Canadian Study ofHealth and Aging16

MARKERS CHARACTERISTICS

0 ● Able to walk without assistance● Able to perform ADLs without assistance

1 ● Bladder incontinence only

2 (Mild frailty) 1 or more of the following (2 ifincontinent):● Needs assistance with mobility or ADLs● Cognitive impairment without dementia● Bowel or bladder incontinence

3 (Moderate/severe frailty) 2 or more of the following (3 ifincontinent):● Totally dependent for transfers● Totally dependent for 1 or more ADLs● Bowel or bladder incontinence● Diagnosis of dementia

ADLs indicates activities of daily living.



and 85 years and older. Similar to the previous stud-ies, frailty served to predict mortality, and was asso-ciated with poor self-ratings of health, more comor-bid illness, and increased social isolation. Datapertaining to specific comorbidities, including cancerdiagnosis, were not reported in this analysis.

Balducci Frailty CriteriaAlthough the above frailty models provide importantprognostic information for the geriatric population atlarge, to the best of our knowledge, data applicable tooncology patients are lacking. Frailty criteria for olderadults with cancer have been proposed, combiningelements of existing definitions. For example, theBalducci criteria are derived in part from observationsreported in a study by Winograd et al assessingpredictors of mortality and institutionalizationamong male patients aged 65 years and older at thePalo Alto Veterans Affairs Medical Center.27 Wino-grad et al identified functional impairment, comor-bidity, and the presence of geriatric syndromes aspredictors. The Balducci frailty criteria (Table 5)supplement these predictors with a criterion of ageolder than 85 years, given the purported high inci-dence of functional and cognitive changes in thisdemographic.28

Mohile et al evaluated the Balducci frailty criteriausing data from the 2003 Medicare Beneficiary Sur-vey.29 With analysis restricted to 12,480 community-dwelling individuals aged 65 years and older, a totalof 2349 patients (19%) were identified with a cancerdiagnosis (excluding skin cancer). Dementia ormemory loss, inability to perform ADLs, and geri-

atric syndromes were assessed by self-report. Subjectssatisfying the Balducci criteria for frailty were ulti-mately found to be more common in the cancergroup versus the noncancer group (80% vs 73%;P � .001). Eight cancer subtypes were also examinedin this multivariate analysis: lung, colon, breast, cer-vical/uterine, prostate, bladder, ovarian, or “other.”Ultimately, a diagnosis of colon cancer (odds ratio[OR], 1.40; 95% confidence interval [95% CI], 1.02-1.93 [P � .04]), breast cancer (OR, 1.32; 95% CI,1.01-1.72 [P � .04]), prostate cancer (OR, 1.45;95% CI, 1.16-1.80 [P � .001]), and bladder cancer(OR, 2.04; 95% CI, 1.21-3.43 [P � .007]) werefound to be significantly associated with frailty. Fur-thermore, it was found that a cancer diagnosis (com-pared with no cancer diagnosis) was associated withpoorer self-rated health (OR, 1.46; 95% CI, 1.29-1.65 [P � .001]).

Using a Geriatric Assessment toPredict Morbidity and MortalityAlthough composite criteria for frailty have beendevised to characterize older adults, the comprehen-sive geriatric assessment contains domains with in-dependent predictive capabilities, including an eval-uation of functional status, comorbid medicalconditions, cognitive function, psychological state,social support, nutritional status, and geriatric syn-dromes. Each of these domains predicts the risk ofmorbidity and mortality in older adults. The domainsof the geriatric assessment are described below.

Functional StatusFunctional status is a strong predictor of morbidityand mortality in the geriatric population.30 Althoughfunctional status can be measured in many differentways, a commonly used tool for evaluating functionalstatus is the ability to complete ADLs and IADLs.ADLs are those basic self-care skills required to main-tain independence in the home, such as the ability tobathe, dress, toilet, transfer, maintain continence, andfeed oneself.31 IADLs are those skills required to main-tain independence in the community, such as the abilityto use the telephone, shop, prepare meals, do house-keeping or laundry, take transportation, take medica-tions, and handle finances.32 Available guidelines rec-ommend assessment of the ability to complete ADLsand IADLs on an annual basis in patients aged older

TABLE 5. Balducci Criteria for Frailty125

FRAILTY CRITERIA CHARACTERISTICS

Age �85 y

ADLs Dependence for 1 or more

Comorbidity 3 or more

Geriatric syndromes One or more of the following:● Delirium● Dementia● Depression● Osteoporosis● Incontinence● Falls● Neglect and abuse● Failure to thrive

ADLs indicates activities of daily living.



than 70 years.33,34 These guidelines may be particularlyrelevant among older adults with cancer because largeepidemiologic studies have reported that a substantialproportion of older patients with cancer need assistancewith the performance of ADLs and IADLs.35 Al-though 70% to 80% of older adults with cancer presentwith an ECOG PS of 0 to 1, greater than half requireassistance with IADLs.36

The need for ADL and IADL assistance is re-ported to be predictive of mortality in the geriatricpopulation.9,37 More recently, need for assistancewith IADLs has been shown to have the same pre-dictive capability among older adults with cancer.38

This relation has also been explored in a disease-specific setting. In a comparison of single-agent orcombination chemotherapy in older adults with non–small cell lung cancer, the need for assistance withIADLs was found to be correlated with poorer sur-vival on multivariate analysis (P � .04).39 The needfor functional assistance may be dependent on stageof disease. An observational study including 1398cancer patients suggested that the need for assistancewith ADLs was more common in hospitalized pa-tients with metastatic cancer compared with patientswith nonmetastatic cancer (31% vs 22%, respectively;P � .0001).40 It appears that among cancer survivors,a greater need for functional assistance persists com-pared with age-matched controls.41

In addition to mortality, assessment of functionalimpairment in older adults with cancer may correlatewith other unique clinical outcomes. For example, theneed for assistance with IADLs appeared to predictpostoperative complications in a series of older adultsundergoing cancer-related surgery (P � .043).42 Simi-larly, in a series of patients with ovarian cancer receivingstandard cytotoxic chemotherapy, the need for func-tional assistance (defined as living at home with assis-tance or living in an institution with medical assistance)was found to be predictive of severe treatment-relatedtoxicity (P � .048).43 Observational data have indicatedthat pain severity in older adults with cancer is corre-lated directly with ADL and IADL disability.44 Func-tional impairment has also been linked more recently topsychological status in older adults with cancer. Theneed for assistance with IADLs has been reported to becorrelated significantly with the incidence of psycholog-ical distress in a population of older adults with cancer(P � .0001).45

ComorbidityThe extent of comorbidity increases with age, andincreasing comorbidity may substantially affect mor-bidity and mortality in older adults with cancer.46,47

This was illustrated in an observational study of17,712 patients with a new primary diagnosis ofcancer, in which survival was found to be inverselyrelated to age and comorbid medical conditions hadan impact on survival independent of cancer stage.46

Similar findings have been noted in disease-specificstudies. For example, in a retrospective cohort studyof 29,733 patients aged 67 years and older withnonmetastatic colorectal cancer, comorbid conditionswere found to be significantly associated with mor-tality (P � .001).48 A similar impact of comorbidityon survival was demonstrated in a study of 936women with early stage breast cancer. Patients with 3or more of 7 selected comorbid medical conditionswere 20 times more likely to die from causes otherthan breast cancer, and patients with severe comor-bidity were uniformly found to have higher mortalityrates compared with patients who had no comorbidconditions.49

Comorbidity may also have a substantial impact ontolerance of anticancer therapy. A prospective assess-ment comparing vinorelbine alone or in combinationwith gemcitabine in patients aged older than 70 yearsidentified an association between a Charlson Comor-bidity Index (CCI) score of 2 or higher and treatmentdiscontinuation.50 Similarly, an assessment of 162patients aged older than 60 years who were treatedwith dose-dense chemotherapy for stage I to IIIbreast cancer suggested that a CCI score of 1 orhigher was associated with an increased risk of grade3 and 4 toxicities. It is interesting to note that 22% ofpatients in this study discontinued therapy prior tothe planned 8 cycles of treatment.51 Certain comor-bidities may also have a specific impact on treatmenttolerance. For example, hypertension may lead to anincreased rate of both trastuzumab-related and an-thracycline-related cardiomyopathy.52,53

Comorbidity also has practical implications forcancer screening. Guidelines from the US PreventiveServices Task Force (USPSTF) and the AmericanCancer Society (ACS) advocate consideration of co-morbid illness and associated prognosis when imple-menting screening for colorectal and breast cancer.54

For example, ACS guidelines indicate that screening



for colorectal cancer may be discontinued in thosepatients with severe comorbidity that would precludetreatment.55 Similarly, USPSTF guidelines for breastcancer suggest that screening may be unnecessary inwomen with comorbid conditions that limit life ex-pectancy.56 An inherent challenge in implementingthese guidelines is the determination of an accurateprognosis.

Psychological State

A link between depression and aging has been estab-lished in several epidemiologic studies. It is estimatedthat 12% to 20% of community-dwelling personsaged 65 years and older experience significant depres-sive symptoms.57-59 Similarly, the link between de-pression and cancer has long been acknowledged.60

Depressive symptoms are particularly prominentamong older, terminally ill, cancer patients. In onereport, increasing age was paralleled by an increasinglikelihood for thoughts of hurting oneself or feeling“better off dead.”61,62 The incidence of suicide has, infact, been noted to be higher among older adults withcancer compared with older adults diagnosed withanother medical illness.63 The use of psychosocialinterventions may be effective in this population, andthe majority of randomized controlled trial (RCT)data assessing psychosocial interventions support thisstrategy.64 These services may be underutilized in anolder population. A study of 326 patients with met-astatic gastrointestinal or lung cancer demonstratedthat 100% of patients aged younger than 40 yearswith depressive symptoms were referred for psycho-social oncology care.65 In contrast, only 22% of pa-tients aged 70 years and older with depressive symp-toms were similarly referred. With regard topharmacotherapy, a review of RCTs for depressioninterventions identified the use of selective serotoninreuptake inhibitors (SSRIs) for 6 weeks and longer asan effective strategy in depressed patients with can-cer.64 However, the potential for interaction betweenanticancer agents (eg, tamoxifen) and SSRIs is animportant consideration.66

Social SupportIn older adults, studies have identified an associationbetween social isolation and mortality.19,67,68 Avail-able social support may also play a role in the diag-nosis and treatment of the older adult with cancer.

A Surveillance, Epidemiology, and End Results(SEER)–Medicare database review assessed womenaged 65 years and older with a diagnosis of breastcancer.69 Those who were unmarried had a higherlikelihood than married women of being diagnosedwith stage II to IV breast cancer versus stage I dis-ease. Furthermore, unmarried women with stage I orII breast cancer were less likely to receive chemother-apy. To the best of our knowledge, limited data existto guide interventions for older adults with cancerwho lack social support. In one study, unmarriedpatients or those living alone were more frequentlyreferred for psychosocial oncology care; however, aspreviously noted, other studies indicate that theseservices are underutilized among older adults withdepressive symptoms.65

Cognitive FunctionDeclines in cognitive function are associated withboth increasing age and an increased risk for all-causemortality.70,71 In the setting of oncology, cognitiveimpairment may be associated with a delay in thediagnosis of cancer. For example, a SEER-Medicarereview examining 17,507 individuals aged 67 yearsand older with invasive colon cancer found that adiagnosis of colon cancer after death (ie, by autopsyor death certificate) was more than twice as likely tooccur in patients with dementia.72 A diagnosis ofcognitive impairment may also alter clinical decision-making in geriatric oncology. A review of the NewMexico Tumor Registry identified a correlation be-tween decreased mental status (assessed via a screen-ing test of cognitive function) and nonreceipt ofdefinitive surgery in 669 cancer patients aged 65 yearsand older.73 The risk of cognitive impairment withanticancer therapy appears to have a strong bearingon patient decision-making as well. In a survey ofpatients aged 60 years and older with cancer, con-gestive heart failure, or chronic obstructive pulmo-nary disease and a limited life expectancy, the major-ity of patients (89%) said they would refuse life-prolonging therapy if it resulted in severe cognitiveimpairment.74

PolypharmacyOlder adults may be particularly susceptible to poly-pharmacy, given the increased number of comorbidi-ties in this population.75 Polypharmacy raises the



likelihood of adverse drug reactions (ADRs) in oldercancer patients, and has been associated with in-creased mortality.76,77 Certain drug classes are im-plicated in a higher frequency of ADRs, includinganticoagulants (specifically warfarin) and benzodiaz-epines.78 Notably, both medications are frequentlyused in the treatment of oncology patients (warfarinfor cancer-related thrombosis, and benzodiazepinesfor the management of anxiety or chemotherapy-related emetogenesis).62,63 Pharmacist-based medica-tion review programs may reduce polypharmacy andthe associated frequency of ADRs, although fol-low-up is needed to determine whether such pro-grams have a substantial impact on clinical outcomein older adults with cancer.79

NutritionSeveral studies in community-dwelling older adultshave identified an association between a low bodymass index (BMI) and an increased risk of mortality(relative to a normal BMI), underscoring the impor-tance of adequate nutrition.80,81 Adequate nutrition isparticularly important in patients with cancer. In ananalysis of 3047 patients enrolled in ECOG proto-cols across multiple malignancies, weight loss wasassociated with lower response rates to chemotherapyand decreased survival.82 The observation of poorerclinical outcome in patients with cancer and concom-itant weight loss has also been made in disease-specific studies of patients with small cell lung cancerand gastrointestinal malignancies.83,84

The nutritional evaluation of older patients should beexpanded in the setting of cancer to include treatment-related considerations. For example, treatment-inducednausea, vomiting, or mucositis can impede oral intake.Furthermore, cancer-related fatigue can hinder the pa-tient’s ability to shop for food and cook, thereby com-promising nutritional status. Various screening toolsassess nutritional status in the older adult, beyond BMIalone. These tools are listed in Table 6.85

Geriatric SyndromesGeriatric syndromes (Table 5) are of particular rele-vance in the cancer patient, and appear to increasewith age.86,87 In the study of Medicare beneficiariesby Mohile et al, a significant association was foundbetween a diagnosis of cancer and a geriatric syn-drome (OR, 1.27; 95% CI, 1.15–1.41).29 These syn-

dromes have, in turn, been linked to multiple adverseclinical outcomes.88 Interventions have been devisedto combat the progression of geriatric syndromes,and data from a randomized study indicate that theseinterventions result in improved functional abilitiesand mental well-being in vulnerable older adults.89

Geriatric Assessment in OncologyPracticeThe geriatric assessment has emerged as a usefuloncology tool. It adds information to more general-ized measures of functional status usually employedin oncology, such as the Karnofsky or ECOG PS.Repetto et al juxtaposed assessment by ECOG PSand by geriatric assessment in a cohort of 363 pa-tients admitted with a diagnosis of a solid or hema-tologic malignancy.90 Elements of the geriatric as-sessment used in this study included demographiccharacteristics, physical function, disability, depres-sion, and cognitive status. Components of the geri-atric assessment and ECOG PS were found to bestrongly associated. For example, the presence ofADL or IADL dependence conferred a 5-fold higherrisk of poor PS (ECOG PS of 2 or higher). However,no association between comorbidity and PS was ob-served in this study. Presumably, these data suggestthat although some domains of the geriatric assess-ment (ie, disability assessment) may serve as a surro-

TABLE 6. Screening Tools to Assess Nutritional Status

TOOL COMPONENTS

Malnutrition UniversalScreening Tool

BMIWeight loss within the past 3-6 moNo nutritional intake in �5 dAcute illness

Short Nutritional AssessmentQuestionnaire

Involuntary weight lossLoss of appetiteUse of tube feeding or supplementaldrinks

Mini Nutritional Assessment Reduction in food intakeWeight lossBMIMobilityAcute psychological distress or acutediseaseNeuropsychological problems: dementiaor depression

Nutritional risk score Undernutrition (BMI, % weight loss,change in food intake)Severity of disease

BMI indicates body mass index.



gate for PS, other components (such as comorbidity)may offer supplementary information that could aug-ment its prognostic value.

A similar tool has been applied in a disease-specificsetting. The geriatric assessment was used to character-ize a series of 84 patients (age 65 years and older) whowere treated for diffuse, large B-cell lymphoma at asingle institution.91 The majority of patients (74%) re-ceived either the combination of cyclophosphamide,doxorubicin, vincristine, and prednisone (CHOP) or aCHOP-like regimen. A total of 42 patients were char-acterized by the geriatric assessment as “fit.” Responserates (93% vs 49%; P � .0001) and median survival (notreached vs 8 months; P � .0001) were found to besuperior in the patients characterized as fit comparedwith those characterized as unfit. Approximately half ofthe patients deemed “unfit” were treated with palliativetherapy, whereas the remainder was treated with cura-tive intent. There did not appear to be a survival dif-ference in either strategy, raising the possibility thatpatients characterized as unfit or frail may derive similarbenefit from palliative approaches. Thus, use of a geri-atric assessment may aid in therapeutic decision-mak-ing, although additional studies are needed.

The geriatric assessment may further aid in pre-dicting treatment toxicity. A prospective analysisconducted at several French institutions over a 12-year interval (1988-2000) included 83 patients aged70 years and older with advanced ovarian carcinomawho were treated with carboplatin and paclitaxel.43 Ageriatric assessment was performed prior to the ini-tiation of therapy. Ultimately, several elements of thegeriatric assessment predicted the occurrence of se-vere toxicity with chemotherapy, including depres-sion (P � .003) and intake of 6 or more nonchemo-therapy medications per day (P � .043).

The geriatric assessment has also been correlatedwith outcomes from surgical resection of solid tumors.Using the Pre-operative Assessment of Cancer in theElderly (PACE) tool, which combines traditional ele-ments of the geriatric assessment with surgery-specificmetrics, a cohort of 460 patients aged older than 70years who underwent cancer-related surgery were eval-uated preoperatively.92 On univariate analysis, ADL orIADL dependence and poor ECOG PS (higher than2) were found to be associated with a longer hospitalstay. On multivariate analysis, components of thePACE tool were found to correlate with postoperativecomplications, including increased fatigue score (ac-

cording to the Brief Fatigue Inventory), ECOG PS(greater than 2), and IADL dependence.

As a logical extension of the prognostic and pre-dictive capabilities of the geriatric assessment, thetool has been applied prospectively to guide the treat-ment of older adults with cancer. In a study of 245patients with cancer who were aged 65 years andolder, a self-administered questionnaire including alldomains of the geriatric assessment was used to guidereferrals to a multidisciplinary team consisting ofsocial workers, psychiatrists, nutritional services,ophthalmologists, otolaryngologists, and rehabilita-tion specialists.93 In addition, referrals were made tocommunity resources (ie, visiting nurses or homehealth aides) on the basis of the questionnaire, andcommunication was initiated with the patient’s pri-mary care provider to assist in managing comorbidmedical conditions that could be exacerbated by thecancer. The self-administered format of the geriatricassessment appeared to be feasible, with patientscompleting the questionnaire within an average of 15minutes. The Cancer and Leukemia Group B(CALGB) is leading an effort to evaluate the feasi-bility of incorporating a primarily self-administeredgeriatric assessment94 into its ongoing protocols(CALGB 360401).95 Efforts such as these may ulti-mately aid in prospectively validating the stratifica-tion of treatment by geriatric assessment result.

Biological Markers of FrailtyEmerging laboratory research has identified molecularmarkers that may be associated with frailty. With in-creasing age, levels of the proinflammatory cytokineinterleukin-6 (IL-6) and C-reactive protein (CRP)demonstrate concomitant increases.96-98 Both appear toplay a role in the development of the frailty criteriaestablished in the CHS, including weight loss. Hub-bard et al assessed frailty characteristics and cytokinelevels in 3 groups of patients: 1) 40 patients housed inan inpatient geriatric ward (median age, 85 years), 2) 40patients attending a day hospital (median age, 83 years),and 3) 30 healthy controls (median age, 23 years).99

Increases in IL-6 and CRP were paralleled by a de-crease in BMI and increasing frailty. In separate studies,increases in IL-6 and CRP have been associated withother domains of the frailty assessment, including poorwalking speed and grip strength.100 In addition, D-dimer and other coagulation markers may play a prog-



nostic role in this population. An assessment of IL-6and D-dimer in 1723 older adults indicated that highercirculating levels of these markers were correlated withan increased risk of mortality.101 Furthermore,D-dimer, factor VIII, and CRP levels in 4735 patientsenrolled in the CHS were found to be associated withclinical frailty.102

Levels of insulin-like growth factor 1 (IGF-1) de-cline with advancing age, and there is some indicationthat this molecular mediator may also play a key role inthe pathogenesis of frailty.103,104 In the WHAS I study,low IGF-1 and high IL-6 levels were associated with amarked increase in walking limitation. Furthermore, inthe subset of patients with low IGF-1 and high IL-6levels, an increase in mortality was observed; at 5 yearsof follow-up, approximately 46% of these patients haddied compared with 23% of patients with high IGF-1and low IL-6 levels.105 Levels of CXC chemokine li-gand 10 (CXCL10) have also been associated with thedevelopment of frailty.106 CXCL10 is a proinflamma-tory chemokine associated with several autoimmunediseases, such as multiple sclerosis and autoimmunethyroiditis.

The molecular markers noted to predict frailty mayalso play a prognostic role in the setting of cancer.For example, elevated CRP has been associated withpoorer survival in patients with localized renal cellcarcinoma, and high D-dimer has been shown topredict poorer survival and disease progression inpatients with metastatic colorectal cancer.107,108 Inaddition, IL-6 may be associated with a poorer prog-nosis in patients with breast cancer, perhaps by up-regulating levels of angiogenic mediators such asvascular endothelial growth factor.109 Elevated IL-6has also been linked to worsened clinical outcome inpatients with advanced non–small cell lung cancer,and may be useful in differentiating hormone-refrac-tory prostate cancer from hormone-sensitive diseaseor benign prostatic conditions.110,111 Ultimately,these molecular markers may guide therapy for theolder adult with cancer. Several of these markers(IL-6, IGF-1, and CXCL10) serve as targets foranticancer agents currently in development.73,112,113

Frailty as a Dynamic Process: ThePotential for InterventionRegardless of the schema used to classify frailty, theimperative is to find ways to identify transitions be-

tween clinical states (ie, from normal to prefrail, or fromprefrail to frail) because these intermediaries offer op-portunities for intervention. Data from the CHS dididentify a greater risk for the development of frailty inthose patients characterized as prefrail.15 For example,patients classified as prefrail at baseline (demonstrating1 or 2 frailty characteristics) were found to have morethan double the risk of becoming frail at 3 to 4 years offollow-up (OR, 2.63; 95% CI, 1.94-3.56 [P � .001]).A prospective exploration of this phenomenon was re-ported by Gill et al, in which 754 community-dwellingindividuals aged 70 years and older were assessed forfrailty at 18-month intervals for a total of 54 months.114

Using the CHS frailty markers, it was noted thatgreater than half of the patients (58%) experienced atransition in frailty status during the evaluation period.Transitions between nonfrail and prefrail states wereconsistent throughout the study period; in contrast, theprobability of going from the frail to the prefrail statedecreased over time; only 2 patients had their statuschange from frail to nonfrail. With extended follow-up,the probability of death in the frail group increased overtime. The likelihood of death in this group of patientscompared with the nonfrail or prefrail groups was 3 to5 times higher, depending on the study period. Thissuggests that frailty represents a dynamic process, un-derscoring the potential for developing strategies toimpede its progression.

Efforts are currently ongoing in the search for waysto reduce the progression of functional decline. A ran-domized trial in frail adults comparing those receivingintensive, home-based physical and occupational ther-apy with those undergoing a less rigorous educationalprogram indicated that the level of disability (defined byADL dependence) declined more rapidly in the pa-tients in the latter group.115 Intervention studies haveaimed to mitigate functional decline in oncology pa-tients through various targeted approaches. One suchstudy identified 641 overweight patients aged 65 yearsand older with breast, colorectal, or prostate cancer.116

These patients were randomly assigned to receive eithera 12-month, home-based program of telephone coun-seling with mailed materials promoting sound physicalactivity and dietary practices or the same interventionafter a 12-month period (representing the “delayed”arm). In comparison with the delayed intervention,immediate use of the intervention led to a lesser declinein physical function, as classified by the Short-Form 36metric.117,118



Emerging data have suggested that pharmacologicinterventions might decrease the risk of frailty. Ananalysis of nonfrail participants (ages 65-79 years) en-rolled in the Women’s Health Initiative ObservationalStudy demonstrated a significant association betweenthe use of low-dose statins for a prolonged period and adecrease in the risk of frailty, as defined by CHS criteria(P � .02).119 The authors postulate that reductions ininflammatory mediators (ie, IL-6 and CRP) associatedwith statin therapy could lead to declines in frailtycharacteristics. These data are particularly provocativein the face of studies that have reported an associationbetween breast cancer and statin use.120-124

ConclusionsSeveral definitions of frailty have been described to date,including those derived from the CHS, the CSHA, andempiric criteria as defined by Balducci et al.15,16,28 Thechallenge for the oncologist is to incorporate elementsfrom these definitions into a framework optimized forthe older adult with cancer. The geriatric assessment

may serve as the framework, because this tool includescomponents with independent prognostic implicationsfor the older cancer patient. Assessing changes in mo-lecular markers associated with frailty, such as IL-6,CRP, and D-dimer, may augment the algorithms usedto stratify risk.

We are beginning to recognize that frailty is adynamic process. This critical observation creates awindow of opportunity for interventions. The geri-atric oncology population would be an ideal one inwhich to study transitions in frailty, in which treat-ment and/or cancer may be associated with eithertransient or permanent functional decline. A betterunderstanding of frailty in an oncology populationmay assist in identifying older adults who are candi-dates for palliative therapy, as well as those patientswho may benefit from standard treatment. Ulti-mately, these data would guide the incorporation offrailty status into algorithms for therapeutic decision-making in cancer care, a much-needed step towarddesigning the best care for vulnerable adults.

References1. He W, Sengupta M, Velkoff V, et al. U.S.

Census Bureau, Current Population Re-ports, P23–209, 65� in the United States:2005. Washington, DC: US GovernmentPrinting Office; 2005.

2. Agingstats.gov. Number of People Age 65and Over, by Age Group, Selected Years1900–2006 and Projected 2010–2050.Available at: http://www.agingstats.gov/Agingstatsdotnet/Main_Site/Data/2008_Documents/Population.aspx. AccessedNovember 23, 2009.

3. Jemal A, Siegel R, Ward E, et al. Cancerstatistics, 2009. CA Cancer J Clin. 2009;59:225-249.

4. Gomez H, Hidalgo M, Casanova L, et al. Riskfactors for treatment-related death in elderlypatientswithaggressivenon-Hodgkin’s lym-phoma: results of a multivariate analysis.J Clin Oncol. 1998;16:2065-2069.

5. Greenlee RT, Murray T, Bolden S, et al.Cancer statistics, 2000. CA Cancer J Clin.2000;50:7-33.

6. Yancik R, Ganz PA, Varricchio CG, et al.Perspectives on comorbidity and cancer inolder patients: approaches to expand theknowledge base. J Clin Oncol. 2001;19:1147-1151.

7. Covinsky KE, Justice AC, Rosenthal GE, etal. Measuring prognosis and case mix inhospitalized elders. The importance offunctional status. J Gen Intern Med. 1997;12:203-208.

8. Klein BEK, Klein R, Knudtson MD, et al.Frailty, morbidity and survival. Arch Ger-ontol Geriatr. 2005;41:141-149.

9. Inouye SK, Peduzzi PN, Robison JT, etal. Importance of functional measures inpredicting mortality among older hos-pitalized patients. JAMA. 1998;279:1187-1193.

10. Walter LC, Brand RJ, Counsell SR, et al.Development and validation of a prognos-tic index for 1-year mortality in olderadults after hospitalization. JAMA. 2001;285:2987-2994.

11. Lee SJ, Lindquist K, Segal MR, et al.Development and validation of a prognos-tic index for 4-year mortality in olderadults. JAMA. 2006;295:801-808.

12. Hutchins LF, Unger JM, Crowley JJ, et al.Underrepresentation of patients 65 yearsof age or older in cancer-treatment trials.N Engl J Med. 1999;341:2061-2067.

13. Extermann M, Muss HB. Conducting clini-cal trials for patients who are frail or elderly.In: Govindan R, ed. American Society ofClinical Oncology 2009 Educational Book.Alexandria, Va; 2009:289-293.

14. [no authors listed]. American MedicalAssociation white paper on elderly health.Report of the Council on Scientific Affairs.Arch Intern Med. 1990;150:2459-2472.

15. Fried LP, Tangen CM, Walston J, et al.Frailty in older adults: evidence for aphenotype. J Gerontol A Biol Sci Med Sci.2001;56:M146-M157.

16. Rockwood K, Howlett SE, MacKnight C,et al. Prevalence, attributes, and out-comesoffitnessandfrailty incommunity-dwelling older adults: report from theCanadian Study of Health and Aging.J Gerontol A Biol Sci Med Sci. 2004;59:1310-1317.

17. Cigolle CT, Ofstedal MB, Tian Z, et al.Comparing models of frailty: the Healthand Retirement Study. J Am Geriatr Soc.2009;57:830-839.

18. Bandeen-Roche K, Xue Q-L, Ferrucci L, etal. Phenotype of frailty: characterizationin the women’s health and aging studies.J Gerontol A Biol Sci Med Sci. 2006;61:262-266.

19. Iwasaki M, Otani T, Sunaga R, et al. Socialnetworks and mortality based on theKomo-Ise cohort study in Japan. Int JEpidemiol. 2002;31:1208-1218.

20. Avila-Funes JA, Helmer C, Amieva H, etal. Frailty among community-dwellingelderly people in France: the three-citystudy. J Gerontol A Biol Sci Med Sci.2008;63:1089-1096.

21. Ensrud KE, Ewing SK, Cawthon PM, et al.A comparison of frailty indexes for theprediction of falls, disability, fractures,and mortality in older men. J Am GeriatrSoc. 2009;57:492-498.

22. Ensrud KE, Ewing SK, Taylor BC, et al.Comparison of 2 frailty indexes for predic-tion of falls, disability, fractures, anddeath in older women. Arch Intern Med.2008;168:382-389.

23. Retornaz F, Monette J, Batist G, et al.Usefulness of frailty markers in the assess-ment of the health and functional status ofolder cancer patients referred for chemo-therapy: a pilot study. J Gerontol A Biol SciMed Sci. 2008;63:518-522.

24. Bylow K, Mohile SG, Stadler WM, et al.Does androgen-deprivation therapy accel-erate the development of frailty in oldermen with prostate cancer? Cancer. 2007;110:2604-2613.



25. Mauras N, Hayes V, Welch S, et al.Testosterone deficiency in young men:marked alterations in whole body proteinkinetics, strength, and adiposity. J ClinEndocrinol Metab. 1998;83:1886-1892.

26. Higano CS. Side effects of androgen depri-vation therapy: monitoring and minimiz-ing toxicity. Urology. 2003;61:32-38.

27. Winograd CH, Gerety MB, Chung M, et al.Screening for frailty: criteria and predic-tors of outcomes. J Am Geriatr Soc. 1991;39:778-784.

28. Balducci L, Extermann M. Management ofthe frail person with advanced cancer. CritRev Oncol Hematol. 2000;33:143-148.

29. Mohile SG, Xian Y, Dale W, et al. Associa-tion of a cancer diagnosis with vulnerabilityand frailty in older Medicare beneficiaries.J Natl Cancer Inst. 2009;101:1206-1215.

30. Reuben DB, Rubenstein LV, Hirsch SH, etal. Value of functional status as a predictorof mortality: results of a prospective study.Am J Med. 1992;93:663-669.

31. Katz S, Ford AB, Moskowitz RW, et al.Studies of illness in the aged: the index ofADL: a standardized measure of biologicaland psychosocial function. JAMA. 1963;185:914-919.

32. Lawton MP, Brody EM. Assessment ofolder people: self-maintaining and instru-mental activities of daily living. Gerontolo-gist. 1969;9:179-186.

33. Rubenstein LV, Calkins DR, Greenfield S,et al. Health status assessment for elderlypatients. Report of the Society of GeneralInternal Medicine Task Force on HealthAssessment. J Am Geriatr Soc. 1989;37:562-569.

34. Applegate WB, Blass JP, Williams TF.Instruments for the functional assessmentof older patients. N Engl J Med. 1990;322:1207-1214.

35. Patel KV, Peek MK, Wong R, et al.Comorbidity and disability in elderlyMexican and Mexican American adults:findings from Mexico and the Southwest-ern United States. J Aging Health. 2006;18:315-329.

36. Extermann M, Hurria A. Comprehensivegeriatric assessment for older patients withcancer. J Clin Oncol. 2007;25:1824-1831.

37. Garman KS, Pieper CF, Seo P, et al.Function in elderly cancer survivors de-pends on comorbidities. J Gerontol A BiolSci Med Sci. 2003;58:M1119-M1124.

38. Wedding U, Rohrig B, Klippstein A, et al.Age, severe comorbidity and functionalimpairment independently contribute topoor survival in cancer patients. J CancerRes Clin Oncol. 2007;133:945-950.

39. Maione P, Perrone F, Gallo C, et al.Pretreatment quality of life and functionalstatus assessment significantly predict sur-vival of elderly patients with advancednon–small-cell lung cancer receiving che-motherapy: a prognostic analysis of themulticenter Italian lung cancer in theelderly. J Clin Oncol. 2005;23:6865-6872.

40. Corsonello A, Pedone C, Carosella L, et al.Health status in older hospitalized pa-tients with cancer or non-neoplasticchronic diseases. BMC Geriatr. 2005;5:10.

41. Sweeney C, Schmitz KH, Lazovich D, et al.Functional limitations in elderly femalecancer survivors. J Natl Cancer Inst. 2006;98:521-529.

42. Audisio RA, Ramesh H, Longo WE, et al.Preoperative assessment of surgical risk inoncogeriatric patients. Oncologist. 2005;10:262-268.

43. Freyer G, Geay J-F, Touzet S, et al.Comprehensive geriatric assessment pre-dicts tolerance to chemotherapy and sur-vival in elderly patients with advancedovarian carcinoma: a GINECO study. AnnOncol. 2005;16:1795-1800.

44. Landi F, Russo A, Liperoti R, et al. Dailypain and functional decline among old-oldadults living in the community: resultsfrom the ilSIRENTE Study. J Pain Symp-tom Manage. 2009;38:350-357.

45. Hurria A, Li D, Hansen K, et al. Distress inolder patients with cancer. J Clin Oncol.2009;27:4346-4351.

46. Piccirillo JF, Tierney RM, Costas I, et al.Prognostic importance of comorbidity in ahospital-based cancer registry. JAMA.2004;291:2441-2447.

47. Yancik R. Cancer burden in the aged: anepidemiologic and demographic over-view. Cancer. 1997;80:1273-1283.

48. Gross CP, Guo Z, McAvay GJ, et al.Multimorbidity and survival in older per-sons with colorectal cancer. J Am GeriatrSoc. 2006;54:1898-1904.

49. Satariano WA, Ragland DR. The effect ofcomorbidity on 3-year survival of womenwith primary breast cancer. Ann InternMed. 1994;120:104-110.

50. Frasci G, Lorusso V, Panza N, et al.Gemcitabine plus vinorelbine versus vi-norelbine alone in elderly patients withadvanced non–small-cell lung cancer.J Clin Oncol. 2000;18:2529-2536.

51. Zauderer M, Patil S, Hurria A. Feasibilityand toxicity of dose-dense adjuvant chemo-therapy in older women with breastcancer. Breast Cancer Res Treat. 2009;117:205-210.

52. Perez EA, Suman VJ, Davidson NE, et al.Cardiac safety analysis of doxorubicin andcyclophosphamide followed by paclitaxelwith or without trastuzumab in the NorthCentral Cancer Treatment Group N9831adjuvant breast cancer trial. J Clin Oncol.2008;26:1231-1238.

53. Suter TM, Procter M, van Veldhuisen DJ,et al. Trastuzumab-associated cardiac ad-verse effects in the herceptin adjuvanttrial. J Clin Oncol. 2007;25:3859-3865.

54. Walter LC, Lewis CL, Barton MB. Screen-ing for colorectal, breast, and cervicalcancer in the elderly: a review of theevidence. Am J Med. 2005;118:1078-1086.

55. Levin B, Lieberman DA, McFarland B, etal. Screening and surveillance for the earlydetection of colorectal cancer and adenom-atous polyps, 2008: a joint guidelinefrom the American Cancer Society, theUS Multi-Society Task Force on Colorec-tal Cancer, and the American Collegeof Radiology. CA Cancer J Clin. 2008;58:130-160.

56. U.S. Department of Health and HumanServices, Agency for Healthcare Researchand Quality. U.S. Preventive Services TaskForce Guidelines for Screening Mammog-raphy. Available at: http://www.ahrq.gov/clinic/USpstf/uspsbrca.htm. Accessed De-cember 3, 2009.

57. Penninx BW, Guralnik JM, Ferrucci L, etal. Depressive symptoms and physicaldecline in community-dwelling older per-sons. JAMA. 1998;279:1720-1726.

58. Broadhead WE, Blazer DG, George LK, etal. Depression, disability days, and dayslost from work in a prospective epidemio-logic survey. JAMA. 1990;264:2524-2528.

59. Unutzer J, Patrick DL, Simon G, et al.Depressive symptoms and the cost ofhealth services in HMO patients aged 65years and older: a 4-year prospectivestudy. JAMA. 1997;277:1618-1623.

60. Goldberg RJ. Management of depressionin the patient with advanced cancer.JAMA. 1981;246:373-376.

61. Breitbart W, Rosenfeld B, Pessin H, et al.Depression, hopelessness, and desire forhastened death in terminally ill patientswith cancer. JAMA. 2000;284:2907-2911.

62. Walker J, Waters RA, Murray G, et al.Better off dead: suicidal thoughts in cancerpatients. J Clin Oncol. 2008;26:4725-4730.

63. Miller M, Mogun H, Azrael D, et al. Cancerand the risk of suicide in older Americans.J Clin Oncol. 2008;26:4720-4724.

64. Dy SM, Lorenz KA, Naeim A, et al.Evidence-based recommendations for can-cer fatigue, anorexia, depression, anddyspnea. J Clin Oncol. 2008;26:3886-3895.

65. Ellis J, Lin J, Walsh A, et al. Predictors ofreferral for specialized psychosocial oncol-ogy care in patients with metastatic cancer:the contributions of age, distress, and mari-tal status. J Clin Oncol. 2009;27:699-705.

66. Stearns V, Johnson MD, Rae JM, et al.Active tamoxifen metabolite plasma con-centrations after coadministration of ta-moxifen and the selective serotonin re-uptake inhibitor paroxetine. J Natl CancerInst. 2003;95:1758-1764.

67. Seeman TE, Kaplan GA, Knudsen L, et al.Social network ties and mortality amongthe elderly in the Alameda County Study.Am J Epidemiol. 1987;126:714-723.

68. Tomaka J, Thompson S, Palacios R. Therelation of social isolation, loneliness, andsocial support to disease outcomes among theelderly. JAgingHealth. 2006;18:359-384.

69. Osborne C, Ostir GV, Du X, et al. Theinfluence of marital status on the stage atdiagnosis, treatment, and survival of olderwomen with breast cancer. Breast CancerRes Treat. 2005;93:41-47.

70. Eagles JM, Beattie JA, Restall DB, et al.Relation between cognitive impairmentand early death in the elderly. BMJ.1990;300:239-240.

71. Wolfson C, Wolfson DB, Asgharian M, etal. A reevaluation of the duration ofsurvival after the onset of dementia. N EnglJ Med. 2001;344:1111-1116.

72. Gupta SK, Lamont EB. Patterns of presenta-tion,diagnosis,andtreatment inolderpatientswith colon cancer and comorbid dementia.JAmGeriatr Soc. 2004;52:1681-1687.

73. Goodwin JS, Hunt WC, Samet JM. Determi-nants of cancer therapy in elderly patients.Cancer. 1993;72:594-601.

74. Fried TR, Bradley EH, Towle VR, et al.Understanding the treatment preferencesof seriously ill patients. N Engl J Med.2002;346:1061-1066.

75. Yancik R, Ershler W, Satariano W, et al.Report of the national institute on agingtask force on comorbidity. J Gerontol ABiol Sci Med Sci. 2007;62:275-280.

76. Flood KL, Carroll MB, Le CV, et al.Polypharmacy in hospitalized older adultcancer patients: experience from a prospec-tive, observational study of an oncology-



acute care for elders unit. Am J GeriatrPharmacother. 2009;7:151-158.

77. Espino DV, Bazaldua OV, Palmer RF, et al.Suboptimal medication use and mortalityin an older adult community-based co-hort: results from the Hispanic EPESEStudy. J Gerontol A Biol Sci Med Sci.2006;61:170-175.

78. Hanlon JT, Pieper CF, Hajjar ER, et al.Incidence and predictors of all and prevent-able adverse drug reactions in frail elderlypersons after hospital stay. J Gerontol ABiol Sci Med Sci. 2006;61:511-515.

79. Vinks TH, Egberts TC, de Lange TM, et al.Pharmacist-based medication review re-duces potential drug-related problems inthe elderly: the SMOG controlled trial.Drugs Aging. 2009;26:123-133.

80. Reynolds MW, Fredman L, Langenberg P,et al. Weight, weight change, mortality ina random sample of older community-dwelling women. J Am Geriatr Soc. 1999;47:1409-1414.

81. Wallace JI, Schwartz RS, LaCroix AZ, et al.Involuntary weight loss in older outpa-tients: incidence and clinical significance.J Am Geriatr Soc. 1995;43:329-337.

82. Dewys WD, Begg C, Lavin PT, et al.Prognostic effect of weight loss prior tochemotherapy in cancer patients. EasternCooperative Oncology Group. Am J Med.1980;69:491-497.

83. Andreyev HJ, Norman AR, Oates J, et al.Why do patients with weight loss have aworse outcome when undergoing chemo-therapy for gastrointestinal malignancies?Eur J Cancer. 1998;34:503-509.

84. Buccheri G, Ferringno D. Importance ofweight loss definition in the prognosticevaluation of non–small-cell lung cancer.Lung Cancer. 2001;34:433-440.

85. Vandewoude M. Nutritional assessment ingeriatric cancer patients. Support Care Can-cer. 2009 Oct 12. [Epub ahead of print].

86. Koroukian SM, Murray P, Madigan E.Comorbidity, disability, and geriatric syn-dromes in elderly cancer patients receiv-ing home health care. J Clin Oncol. 2006;24:2304-2310.

87. Flood KL, Carroll MB, Le CV, et al. Geriatricsyndromes in elderly patients admitted to anoncology-acute care for elders unit. J ClinOncol. 2006;24:2298-2303.

88. AnpalahanM,GibsonSJ.Geriatric syndromesas predictors of adverse outcomes of hospital-ization. InternMedJ. 2008;38:16-23.

89. Phelan EA, Balderson B, Levine M, et al.Delivering effective primary care to olderadults: a randomized, controlled trial ofthe senior resource team at group healthcooperative. J Am Geriatr Soc. 2007;55:1748-1756.

90. Repetto L, Fratino L, Audisio RA, et al.Comprehensive geriatric assessment addsinformation to Eastern Cooperative Oncol-ogy Group performance status in elderlycancer patients: an Italian Group for Geri-atric Oncology Study. J Clin Oncol. 2002;20:494-502.

91. Tucci A, Ferrari S, Bottelli C, et al. Acomprehensive geriatric assessment is moreeffective than clinical judgment to identifyelderly diffuse large cell lymphoma patientswho benefit from aggressive therapy. Can-cer. 2009;115:4547-4453.

92. PACE participants, Audisio RA, Pope D, etal. Shall we operate? Preoperative assess-ment in elderly cancer patients (PACE)

can help. A SIOG surgical task forceprospective study. Crit Rev Oncol Hematol.2008;65:156-163.

93. Hurria A, Lichtman SM, Gardes J, et al.Identifying vulnerable older adults withcancer: integrating geriatric assessmentinto oncology practice. J Am Geriatr Soc.2007;55:1604-1608.

94. Hurria A, Gupta S, Zauderer M, et al.Developing a cancer-specific geriatric as-sessment: a feasibility study. Cancer. 2005;104:1998-2005.

95. Cohen HJ, Muss HB. The cancer and leuke-mia group B cancer in the elderly commit-tee: addressing a major cancer need. ClinCancer Res. 2006;12:3606s-3611s.

96. Abbatecola AM, Ferrucci L, Grella R, et al.Diverse effect of inflammatory markers oninsulin resistance and insulin-resistancesyndrome in the elderly. J Am Geriatr Soc.2004;52:399-404.

97. Fagiolo U, Cossarizza A, Scala E, et al.Increased cytokine production in mono-nuclear cells of healthy elderly people. EurJ Immunol. 1993;23:2375-2378.

98. Roubenoff R, Harris TB, Abad LW, et al.Monocyte cytokine production in an el-derly population: effect of age and inflam-mation. J Gerontol A Biol Sci Med Sci.1998;53:M20-M26.

99. Hubbard RE, O’Mahony MS, Calver BL, etal. Nutrition, inflammation, and leptinlevels in aging and frailty. J Am GeriatrSoc. 2008;56:279-284.

100. Taaffe DR, Harris TB, Ferrucci L, et al.Cross-sectional and prospective relation-ships of interleukin-6 and C-reactive pro-tein with physical performance in elderlypersons: MacArthur studies of successfulaging. J Gerontol A Biol Sci Med Sci.2000;55:M709-M715.

101. Cohen HJ, Harris T, Pieper CF. Coagula-tion and activation of inflammatory path-ways in the development of functionaldecline and mortality in the elderly. Am JMed. 2003;114:180-187.

102. Walston J, McBurnie MA, Newman A, etal. Frailty and activation of the inflamma-tion and coagulation systems with andwithout clinical comorbidities: results fromthe Cardiovascular Health Study. ArchIntern Med. 2002;162:2333-2341.

103. Papadakis MA, Grady D, Tierney MJ, et al.Insulin-like growth factor 1 and functionalstatus in healthy older men. J Am GeriatrSoc. 1995;43:12.

104. Ershler WB, Keller ET. Age-associatedincreased interleukin-6 gene expression,late-life diseases, and frailty. Annu RevMed. 2000;51:245-270.

105. Cappola AR, Xue Q-L, Ferrucci L, et al.Insulin-like growth factor I and interleu-kin-6 contribute synergistically to disabilityand mortality in older women. J Clin Endo-crinol Metab. 2003;88:2019-2025.

106. Qu T, Yang H, Walston JD, et al. Upregu-lated monocytic expression of CXC chemo-kine ligand 10 (CXCL-10) and its relation-ship with serum interleukin-6 levels inthe syndrome of frailty. Cytokine. 2009;46:319-324.

107. Blackwell K, Hurwitz H, Lieberman G, et al.Circulating D-dimer levels are better predic-tors of overall survival and disease progres-sion than carcinoembryonic antigen levelsin patients with metastatic colorectal carci-noma. Cancer. 2004;101:77-82.

108. Ramsey S, Lamb GW, Aitchison M, et al.Evaluation of an inflammation-based prog-

nostic score in patients with metastatic renalcancer. Cancer. 2007;109:205-212.

109. Caruso C, Lio D, Cavallone L, et al. Aging,longevity, inflammation, and cancer. AnnN Y Acad Sci. 2004;1028:1-13.

110. Drachenberg DE, Elgamal AA, Row-botham R, et al. Circulating levels ofinterleukin-6 in patients with hormonerefractory prostate cancer. Prostate. 1999;41:127-133.

111. Songur N, Kuru B, Kalkan F, et al. Seruminterleukin-6 levels correlate with malnu-trition and survival in patients with ad-vanced non–small cell lung cancer. Tu-mori. 2004;90:196-200.

112. Schagen SB, van Dam FS, Muller MJ, et al.Cognitive deficits after postoperative adju-vant chemotherapy for breast carcinoma.Cancer. 1999;85:640-650.

113. Schagen SB, Muller MJ, Boogerd W, et al.Cognitive dysfunction and chemotherapy:neuropsychological findings in perspec-tive. Clin Breast Cancer. 2002;3(suppl3):S100-S108.

114. Gill TM, Gahbauer EA, Allore HG, et al.Transitions between frailty states amongcommunity-living older persons. Arch In-tern Med. 2006;166:418-423.

115. Gill TM, Baker DI, Gottschalk M, et al. Aprogramtopreventfunctionaldeclineinphysi-cally frail, elderly persons who live at home.NEngl JMed. 2002;347:1068-1074.

116. Morey MC, Snyder DC, Sloane R, et al.Effects of home-based diet and exercise onfunctional outcomes among older, over-weight long-term cancer survivors: RE-NEW: a randomized controlled trial.JAMA. 2009;301:1883-1891.

117. Ware JE Jr, Sherbourne CD. The MOS36-item short-form health survey (SF-36).I. Conceptual framework and item selec-tion. Med Care. 1992;30:473-483.

118. McHorney CA, Ware JE Jr, Raczek AE.The MOS 36-Item Short-Form Health Sur-vey (SF-36): II. Psychometric and clinicaltests of validity in measuring physical andmental health constructs. Med Care. 1993;31:247-263.

119. LaCroix AZ, Gray SL, Aragaki A, et al.Statin use and incident frailty in womenaged 65 years or older: prospective find-ings from the Women’s Health InitiativeObservational Study. J Gerontol A Biol SciMed Sci. 2008;63:369-375.

120. Garwood E, Kumar A, Baehner F, et al.Fluvastatin reduces proliferation and in-creases apoptosis in women with high gradebreast cancer. Breast Cancer Res Treat. 2009Aug 29. [Epub ahead of print].

121. Pocobelli G, Newcomb PA, Trentham-Dietz A, et al. Statin use and risk of breastcancer. Cancer. 2008;112:27-33.

122. Kwan M, Habel L, Flick E, et al. Post-diagnosis statin use and breast cancerrecurrence in a prospective cohort study ofearly stage breast cancer survivors. BreastCancer Res Treat. 2008;109:573-579.

123. Boudreau DM, Yu O, Miglioretti DL, et al.Statin use and breast cancer risk in a largepopulation-based setting. Cancer EpidemiolBiomarkers Prev. 2007;16:416-421.

124. Kuoppala J, Lamminpaa A, Pukkala E.Statins and cancer: a systematic review andmeta-analysis. Eur J Cancer. 2008;44:2122-2132.

125. Balducci L, Extermann M. Managementof cancer in the older person: a practicalapproach. Oncologist. 2000;5:224-237.

