[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Figure 1.  Overall Survival Probability After Cancer Diagnosis Among 7257 Women in the WHI Study by Prediagnostic Frailty Status
Overall Survival Probability After Cancer Diagnosis Among 7257 Women in the WHI Study by Prediagnostic Frailty Status

WHI indicates Women’s Health Initiative.

Figure 2.  Rate of Increase in Physical Frailty From Before to After Cancer Diagnosis Among 7257 Women in the WHI Study
Rate of Increase in Physical Frailty From Before to After Cancer Diagnosis Among 7257 Women in the WHI Study

Physical frailty assessments are based on the frailty score at the 3-year Women’s Health Initiative (WHI) study visit and subsequent updates to the physical function component of the frailty score. Analyses are restricted to women diagnosed as having cancer. Shown is the change in frailty score in relation to cancer diagnosis from a linear mixed-effects model with a piecewise linear spline (knot at date of cancer diagnosis) and a random intercept and slope for each individual.

Figure 3.  Time-Varying Frailty Status and Mortality After Cancer Diagnosis Among 7257 Women in the Women’s Health Initiative (WHI) Study
Time-Varying Frailty Status and Mortality After Cancer Diagnosis Among 7257 Women in the Women’s Health Initiative (WHI) Study

Frailty scores are a time-varying exposure, with nonfrail as the reference. All Cox proportional hazards models adjusted for the following covariates at the 3-year WHI visit: race/ethnicity, body mass index, pack-years of smoking, educational attainment, Charlson Comorbidity Index score, and any family history of cancer. P for trend was calculated for frailty categories as an ordinal variable. All analyses were restricted to women without a prior cancer diagnosis at the 3-year WHI visit who were followed up for mortality beginning from the date of their first cancer diagnosis through March 2018. HR indicates hazard ratio.

Table 1.  Association of Prediagnostic Frailty With Mortality After Cancer Diagnosis Among 7257 Women in the WHI Studya
Association of Prediagnostic Frailty With Mortality After Cancer Diagnosis Among 7257 Women in the WHI Studya
Table 2.  Association of Prediagnostic Change in Frailty With Mortality After Cancer Diagnosis Among 7257 Women in the WHI Studya
Association of Prediagnostic Change in Frailty With Mortality After Cancer Diagnosis Among 7257 Women in the WHI Studya
1.
Bluethmann  SM, Mariotto  AB, Rowland  JH.  Anticipating the “silver tsunami”: prevalence trajectories and comorbidity burden among older cancer survivors in the United States.   Cancer Epidemiol Biomarkers Prev. 2016;25(7):1029-1036. doi:10.1158/1055-9965.EPI-16-0133 PubMedGoogle Scholar
2.
Clegg  A, Young  J, Iliffe  S, Rikkert  MO, Rockwood  K.  Frailty in elderly people.   Lancet. 2013;381(9868):752-762. doi:10.1016/S0140-6736(12)62167-9 PubMedGoogle Scholar
3.
Ethun  CG, Bilen  MA, Jani  AB, Maithel  SK, Ogan  K, Master  VA.  Frailty and cancer: implications for oncology surgery, medical oncology, and radiation oncology.   CA Cancer J Clin. 2017;67(5):362-377. doi:10.3322/caac.21406 PubMedGoogle Scholar
4.
Wildiers  H, Heeren  P, Puts  M,  et al.  International Society of Geriatric Oncology consensus on geriatric assessment in older patients with cancer.   J Clin Oncol. 2014;32(24):2595-2603. doi:10.1200/JCO.2013.54.8347 PubMedGoogle Scholar
5.
Woods  NF, LaCroix  AZ, Gray  SL,  et al; Women’s Health Initiative.  Frailty: emergence and consequences in women aged 65 and older in the Women’s Health Initiative Observational Study.   J Am Geriatr Soc. 2005;53(8):1321-1330. doi:10.1111/j.1532-5415.2005.53405.x PubMedGoogle Scholar
6.
Fried  LP, Tangen  CM, Walston  J,  et al; Cardiovascular Health Study Collaborative Research Group.  Frailty in older adults: evidence for a phenotype.   J Gerontol A Biol Sci Med Sci. 2001;56(3):M146-M156. doi:10.1093/gerona/56.3.M146 PubMedGoogle Scholar
7.
 Erratum.   J Am Geriatr Soc. 2017;65(7):1631-1632. doi:10.1111/jgs.15006 PubMedGoogle Scholar
8.
Brundle  C, Heaven  A, Brown  L,  et al.  Convergent validity of the electronic frailty index.   Age Ageing. 2019;48(1):152-156. doi:10.1093/ageing/afy162 PubMedGoogle Scholar
9.
Clegg  A, Bates  C, Young  J,  et al.  Development and validation of an electronic frailty index using routine primary care electronic health record data.   Age Ageing. 2016;45(3):353-360. doi:10.1093/ageing/afw039 PubMedGoogle Scholar
10.
Rockwood  K, Mitnitski  A.  Frailty, fitness, and the mathematics of deficit accumulation.   Rev Clin Gerontol. 2007;17(1):1-12. doi:10.1017/S0959259807002353 Google Scholar
11.
Gilbert  T, Neuburger  J, Kraindler  J,  et al.  Development and validation of a Hospital Frailty Risk Score focusing on older people in acute care settings using electronic hospital records: an observational study.   Lancet. 2018;391(10132):1775-1782. doi:10.1016/S0140-6736(18)30668-8 PubMedGoogle Scholar
12.
Handforth  C, Clegg  A, Young  C,  et al.  The prevalence and outcomes of frailty in older cancer patients: a systematic review.   Ann Oncol. 2015;26(6):1091-1101. doi:10.1093/annonc/mdu540 PubMedGoogle Scholar
13.
Balducci  L, Fossa  SD.  Rehabilitation of older cancer patients.   Acta Oncol. 2013;52(2):233-238. doi:10.3109/0284186X.2012.744142 PubMedGoogle Scholar
14.
Mohile  SG, Xian  Y, Dale  W,  et al.  Association of a cancer diagnosis with vulnerability and frailty in older Medicare beneficiaries.   J Natl Cancer Inst. 2009;101(17):1206-1215. doi:10.1093/jnci/djp239 PubMedGoogle Scholar
15.
Pérez-Zepeda  MU, Cárdenas-Cárdenas  E, Cesari  M, Navarrete-Reyes  AP, Gutiérrez-Robledo  LM.  Cancer and frailty in older adults: a nested case-control study of the Mexican Health and Aging Study.   J Cancer Surviv. 2016;10(4):736-742. doi:10.1007/s11764-016-0519-6 PubMedGoogle Scholar
16.
Women’s Health Initiative Study Group.  Design of the Women’s Health Initiative clinical trial and observational study.   Control Clin Trials. 1998;19(1):61-109. doi:10.1016/S0197-2456(97)00078-0 PubMedGoogle Scholar
17.
Anderson  GL, Manson  J, Wallace  R,  et al.  Implementation of the Women’s Health Initiative study design.   Ann Epidemiol. 2003;13(9 suppl):S5-S17. doi:10.1016/S1047-2797(03)00043-7 PubMedGoogle Scholar
18.
McTiernan  A, Rossouw  J, Manson  J,  et al  Informed consent in the Women’s Health Initiative clinical trial and observational study.   J Womens Health. 1995;4(5):519-529. doi:10.1089/jwh.1995.4.519 Google Scholar
19.
Hays  RD, Morales  LS.  The RAND-36 measure of health-related quality of life.   Ann Med. 2001;33(5):350-357. doi:10.3109/07853890109002089 PubMedGoogle Scholar
20.
Ware  JE  Jr.  SF-36 health survey update.   Spine (Phila Pa 1976). 2000;25(24):3130-3139. doi:10.1097/00007632-200012150-00008 PubMedGoogle Scholar
21.
Zaslavsky  O, Zelber-Sagi  S, Gray  SL,  et al.  Comparison of frailty phenotypes for prediction of mortality, incident falls, and hip fracture in older women.   J Am Geriatr Soc. 2016;64(9):1858-1862. doi:10.1111/jgs.14233 PubMedGoogle Scholar
22.
National Cancer Institute. Surveillance, Epidemiology, and End Results Program. Accessed February 17, 2019. https://seer.cancer.gov/
23.
Stampfer  MJ, Willett  WC, Speizer  FE,  et al.  Test of the National Death Index.   Am J Epidemiol. 1984;119(5):837-839. doi:10.1093/oxfordjournals.aje.a113804 PubMedGoogle Scholar
24.
Rockwood  K, Howlett  SE, MacKnight  C,  et al.  Prevalence, attributes, and outcomes of fitness and frailty in community-dwelling older adults: report from the Canadian Study of Health and Aging.   J Gerontol A Biol Sci Med Sci. 2004;59(12):1310-1317. doi:10.1093/gerona/59.12.1310 PubMedGoogle Scholar
25.
Parker  SG, McCue  P, Phelps  K,  et al.  What is Comprehensive Geriatric Assessment (CGA)? an umbrella review.   Age Ageing. 2018;47(1):149-155. doi:10.1093/ageing/afx166 PubMedGoogle Scholar
26.
Li  CM, Chen  CY, Li  CY, Wang  WD, Wu  SC.  The effectiveness of a comprehensive geriatric assessment intervention program for frailty in community-dwelling older people: a randomized, controlled trial.   Arch Gerontol Geriatr. 2010;50(suppl 1):S39-S42. doi:10.1016/S0167-4943(10)70011-X PubMedGoogle Scholar
27.
Rockwood  K, Andrew  M, Mitnitski  A.  A comparison of two approaches to measuring frailty in elderly people.   J Gerontol A Biol Sci Med Sci. 2007;62(7):738-743. doi:10.1093/gerona/62.7.738 PubMedGoogle Scholar
28.
Petrick  JL, Foraker  RE, Kucharska-Newton  AM,  et al.  Trajectory of overall health from self-report and factors contributing to health declines among cancer survivors.   Cancer Causes Control. 2014;25(9):1179-1186. doi:10.1007/s10552-014-0421-3 PubMedGoogle Scholar
29.
Petrick  JL, Reeve  BB, Kucharska-Newton  AM,  et al.  Functional status declines among cancer survivors: trajectory and contributing factors.   J Geriatr Oncol. 2014;5(4):359-367. doi:10.1016/j.jgo.2014.06.002 PubMedGoogle Scholar
30.
Hodes  RJ, Sierra  F, Austad  SN,  et al.  Disease drivers of aging.   Ann N Y Acad Sci. 2016;1386(1):45-68. doi:10.1111/nyas.13299 PubMedGoogle Scholar
31.
Kidd  T, Mold  F, Jones  C,  et al.  What are the most effective interventions to improve physical performance in pre-frail and frail adults? a systematic review of randomised control trials.   BMC Geriatr. 2019;19(1):184. doi:10.1186/s12877-019-1196-x PubMedGoogle Scholar
32.
Hernández Morante  JJ, Gómez Martínez  C, Morillas-Ruiz  JM.  Dietary factors associated with frailty in old adults: a review of nutritional interventions to prevent frailty development.   Nutrients. 2019;11(1):E102. doi:10.3390/nu11010102 PubMedGoogle Scholar
33.
Daum  CW, Cochrane  SK, Fitzgerald  JD, Johnson  L, Buford  TW.  Exercise interventions for preserving physical function among cancer survivors in middle to late life.   J Frailty Aging. 2016;5(4):214-224.PubMedGoogle Scholar
34.
Arem  H, Moore  SC, Park  Y,  et al.  Physical activity and cancer-specific mortality in the NIH-AARP Diet and Health Study cohort.   Int J Cancer. 2014;135(2):423-431. doi:10.1002/ijc.28659 PubMedGoogle Scholar
35.
Borch  KB, Braaten  T, Lund  E, Weiderpass  E.  Physical activity before and after breast cancer diagnosis and survival: the Norwegian Women and Cancer cohort study.   BMC Cancer. 2015;15:967. doi:10.1186/s12885-015-1971-9 PubMedGoogle Scholar
36.
Demark-Wahnefried  W.  Cancer survival: time to get moving? data accumulate suggesting a link between physical activity and cancer survival.   J Clin Oncol. 2006;24(22):3517-3518. doi:10.1200/JCO.2006.06.6548 PubMedGoogle Scholar
37.
O’Donovan  G, Lee  IM, Hamer  M, Stamatakis  E.  Association of “weekend warrior” and other leisure time physical activity patterns with risks for all-cause, cardiovascular disease, and cancer mortality.   JAMA Intern Med. 2017;177(3):335-342. doi:10.1001/jamainternmed.2016.8014 PubMedGoogle Scholar
38.
Parekh  N, Lin  Y, Craft  LL, Vadiveloo  M, Lu-Yao  GL.  Longitudinal associations of leisure-time physical activity and cancer mortality in the Third National Health and Nutrition Examination Survey (1986-2006).   J Obes. 2012;2012:518358. doi:10.1155/2012/518358 PubMedGoogle Scholar
39.
Vainshelboim  B, Müller  J, Lima  RM,  et al.  Cardiorespiratory fitness, physical activity and cancer mortality in men.   Prev Med. 2017;100:89-94. doi:10.1016/j.ypmed.2017.04.014 PubMedGoogle Scholar
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Views 784
    Citations 0
    Original Investigation
    Geriatrics
    September 14, 2020

    Association of Prediagnostic Frailty, Change in Frailty Status, and Mortality After Cancer Diagnosis in the Women’s Health Initiative

    Author Affiliations
    • 1Division of Research, Kaiser Permanente Northern California, Oakland
    • 2Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, Washington
    • 3Division of Hematology, Department of Internal Medicine, The Ohio State University, Columbus
    • 4Division of Cancer Prevention and Control, Department of Internal Medicine, The Ohio State University, Columbus
    • 5School of Nursing, University of Washington, Seattle
    • 6Department of Epidemiology, College of Public Health, The University of Iowa, Iowa City
    JAMA Netw Open. 2020;3(9):e2016747. doi:10.1001/jamanetworkopen.2020.16747
    Key Points

    Question  Are prediagnostic physical frailty and change in frailty status over time associated with mortality after a cancer diagnosis?

    Findings  In this cohort study of 7257 postmenopausal women with cancer, sustained and worsening frailty before a cancer diagnosis was associated with a more than 20% increased risk of mortality after cancer diagnosis. Furthermore, the rate of decline in physical function accelerated after cancer diagnosis.

    Meaning  These data suggest that frailty assessment in patients with cancer could provide valuable prognostic information and perhaps prompt interventions to reduce or preempt worsening of physical frailty after a cancer diagnosis.

    Abstract

    Importance  Understanding changes in frailty in relation to cancer diagnosis can inform optimal selection of cancer treatments and survivorship care.

    Objective  To investigate associations of prediagnostic frailty and change in frailty status with mortality after a cancer diagnosis.

    Design, Setting, and Participants  This multicenter, prospective cohort study included 7257 community-dwelling, postmenopausal women in the United States who had frailty assessed at the Women’s Health Initiative (WHI) enrollment (1993-1998) and the 3-year visit who were subsequently diagnosed as having invasive cancer. The data were analyzed from January 7, 2019, to June, 8, 2020.

    Exposure  Frailty scores were defined from validated questionnaire items conceptually aligned with the Fried frailty phenotype, including at least 3 of the following characteristics: self-reported unintentional weight loss, exhaustion, low physical activity, and muscle weakness or impaired walking. Physical function components of the frailty score were updated a median of 10 (range, 1-18) times.

    Main Outcomes and Measures  Using multivariable-adjusted Cox proportional hazards models, this study examined associations of prediagnostic frailty (at the 3-year visit, before cancer diagnosis) and prediagnostic changes in frailty (from enrollment to the 3-year visit) with mortality. Women were followed up beginning from cancer diagnosis for mortality outcomes through March 2018. In linear mixed-effects models with frailty scores as a function of time since cancer diagnosis, this study evaluated whether the time slope, ie, the rate of change in frailty score, increased after cancer diagnosis.

    Results  This study included 7257 women in the WHI cohort who completed frailty assessments at enrollment and the 3-year WHI visit before cancer diagnosis and subsequently developed cancer. Cancer cases included 2644 breast cancers (36%), 822 lung cancers (11%), 691 colorectal cancers (10%), 445 endometrial cancers (6%), and 286 ovarian cancers (4%). At the 3-year visit, prior to cancer diagnosis, the mean (SD) age was 63 (7) years, and 1161 of 7257 (16%) of participating women met criteria for frailty; 2129 of 7257 (29%) were prefrail, and 3967 of 7257 (55%) were nonfrail. Over a median follow-up of 5.8 years after cancer diagnosis (range, 1 day to 19.9 years), 3056 women died. After multivariable adjustment, women who were frail (vs nonfrail) before cancer diagnosis had an increased risk of mortality after cancer diagnosis (hazard ratio [HR], 1.40; 95% CI, 1.26-1.55; P for trend <.001). Sustained frailty (21% [1537 of 7257] of women) or worsening frailty (22% [1578 of 7257]) vs being consistently nonfrail (45% [3266 of 7257]) before cancer diagnosis increased the risk of mortality after cancer diagnosis (HR, 1.25; 95% CI, 1.14-1.38 and 1.22; 95% CI, 1.11-1.34, respectively; P for trend <.001). In linear mixed-effects models, the rate of increase in physical frailty over time was statistically significantly higher after cancer diagnosis.

    Conclusions and Relevance  Sustained and worsening frailty before cancer diagnosis was associated with an increased risk of mortality after cancer diagnosis in postmenopausal women. Furthermore, the rate of decline in physical function accelerated after cancer diagnosis. Frailty assessment could provide valuable information and perhaps prompt interventions to reduce and preempt worsening of physical frailty after cancer diagnosis.

    Introduction

    There are 16 million cancer survivors in the United States, and 60% are 65 years or older; by 2040, the number of survivors will increase to 26 million, with 73% being 65 years and older.1 Given this demographic shift, interest is growing in how age and age-related comorbidities alter cancer survival and how cancer and its treatments change trajectories of healthy aging. Frailty, diminished physiologic reserve that results in decreased resiliency and adaptive capacity and increased vulnerability to stressors,2 is an aging syndrome prevalent among older adults that is associated with morbidity and mortality.3,4

    Two widely accepted frailty paradigms are the frailty phenotype and the cumulative deficit model. Although these assessments overlap to some degree, each has advantages. The frailty phenotype uses the Fried criteria to describe a clinically reproducible syndrome that includes weight loss, exhaustion, low physical activity, slowness, and weakness.5-7 The Fried frailty phenotype can be evaluated at first contact to stratify patients as nonfrail, prefrail, or frail. The cumulative deficit model is assessed by continuous scores (eg, the Rockwood index) that tally accumulated physical and cognitive deficits arising from diseases and disabilities.8-11 Such indexes provide granular assessment that is practical in retrospective medical record cohorts.

    Although the prevalence of frailty varies by the definition used, more than one-half of older patients with cancer are frail or prefrail at cancer diagnosis; these patients are thought to be at increased risk of chemotherapy intolerance, postoperative complications, and mortality.12 In addition to preexisting frailty, cancer and its treatments present additional stressors that can accelerate aging and worsen frailty status.5,7,13-15 Understanding the association of frailty and change in frailty status with cancer diagnosis and subsequent mortality is important for optimal selection of cancer treatments and survivorship care.3,4 Despite the dynamic nature of frailty, to our knowledge, no prior study has examined whether changes in functional decline over time are associated with cancer mortality.

    To fill this gap, the present study included 7257 participants in the Women’s Health Initiative (WHI) who developed invasive cancer. At enrollment and 3 years after enrollment, frailty scores were defined from validated questionnaire items conceptually aligned with the Fried frailty phenotype, including at least 3 of the following characteristics: self-reported unintentional weight loss, exhaustion, low physical activity, and muscle weakness or impaired walking. Women exhibiting up to 2 of these characteristics were considered prefrail.5-7 We then examined associations of prediagnostic frailty (at the 3-year visit, before cancer diagnosis) and prediagnostic changes in frailty (from enrollment to the 3-year visit) with mortality. Women were followed up beginning from cancer diagnosis for mortality outcomes through March 2018. Secondarily, we examined whether the rate of increase in frailty changed after cancer diagnosis.

    Methods
    Study Population

    The design and conduct of the WHI has been previously described.16,17 In brief, the WHI observational study (N = 93 676) included postmenopausal women aged 50 to 79 years from 40 US clinical centers between 1993 and 1998. At enrollment, self-administered questionnaires were used to collect information on demographic characteristics; medical, reproductive, and family history; and lifestyle factors, including smoking and physical activity. Written informed consent materials were reviewed by study investigators, local institutional review boards, the WHI data and safety monitoring board, the WHI policy advisory committee, the National Institutes of Health, the Institute of Medicine, focus groups of postmenopausal women, and media.18 Written informed consent was obtained from all participants, and this study was approved by each institution’s institutional review board. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

    For this prospective cohort study, our main exposure was defined as frailty assessed at the 3-year WHI visit (so as to examine change in frailty). Eligible participants were enrolled in the WHI (N = 93 676) and had a first diagnosis of cancer after the 3-year visit (n = 11 913). Women were excluded from this set of 11 913 if they reported a history of cancer before WHI enrollment (1993-1998), were missing covariate data, or were missing frailty status at enrollment or at the 3-year WHI visit (n = 4656). Therefore, after exclusions, this multicenter study included 7257 US community-dwelling, postmenopausal women with frailty assessed at 2 time points before diagnosis of invasive cancer. Women included vs women excluded were similar with respect to age, race/ethnicity, and body mass index (calculated as weight in kilograms divided by height in meters squared). Data analysis was conducted from January 7, 2019, to June, 8, 2020.

    Exposures

    The main exposure was a frailty score, which was defined as in prior publications5-7 using validated questionnaire items to operationalize components of the Fried frailty phenotype. At enrollment, these components included self-reported unintentional weight loss of at least 15 lb (6.8 kg [to convert lbs to kg, multiply by 0.45]) in the past 6 months, self-reported low physical activity (<2.5 metabolic equivalent task hours per week), and muscle weakness or impaired walking and exhaustion (measured using the 36-Item Short Form Health Survey [SF-36] physical function [<70] and vitality [<55] scores).19-21 These measures were repeated at the 3-year WHI visit, with unintentional weight loss redefined based on measured weight loss at the clinic visit exceeding 5% since enrollment and no self-reported indication of an intention to lose at least 5 lb in the past 2 years. As in prior publications,5-7 for each measure (except unintentional weight loss), a frailty component was classified as present if the participant had a score in the lowest quartile of the distribution for that component. To align the scoring with the Fried criteria, poor physical function was scored as 2 points because the SF-36 measures both muscle strength and walking ability components. The number of frailty components present was summed, yielding a range of 0 to 5. A frailty cut point of at least 3 was used, with women scoring 1 to 2 considered prefrail and those scoring 0 considered robust or nonfrail.5-7 Change in frailty was computed as a continuous change in score from enrollment to the 3-year WHI visit and as a categorical variable indicating an increase or a decrease or that the woman scored as frail or nonfrail both times.

    The eFigure in the Supplement shows the time line for frailty data collection in the WHI and the analytic time line for our analysis. All women had frailty scores measured at 2 time points before cancer diagnosis (WHI enrollment and the 3-year visit). In addition, most women participated in the WHI extension studies (6318 [87%] in 2005-2010 and 5157 [71%] in 2010-2015) and thus had updated measurements available for time-varying analyses. The physical function components of the frailty score assessed using the SF-36 were updated annually, whereas the energy and fatigue components were updated only once, in 2012. Weight change and physical activity were updated only at enrollment and the 3-year visit. Overall, physical function components of the frailty score were updated a median of 10 (range, 1-18) times; among women participating in the extension studies, the median number of frailty score updates was 12 (range, 1-18).

    Outcomes

    Annually, newly reported cancer diagnoses were documented with medical records and confirmed and coded at the WHI Clinical Coordinating Center.22 Deaths were ascertained through numerous sources (eg, reports from relatives, returned mail, and obituary notices) and documented with medical records or death certificates. Serial National Death Index queries through March 2018 provided death certificate information regardless of reconsent status, resulting in overall survival information that is 98% complete.23

    Covariates

    Using a standardized approach, trained research staff measured participants’ height and weight, from which we computed body mass index. Via questionnaires, participants reported total physical activity, pack-years of smoking, educational attainment, family income, diagnosis of or treatment for diabetes, and any family history of cancer. In addition, we calculated the Charlson Comorbidity Index score for each participant at the 3-year visit and the prevalence of individual chronic conditions associated with frailty, including history of rheumatoid arthritis, history of congestive heart failure, emphysema, chronic obstructive pulmonary disease (COPD), liver disease, and chronic kidney disease.

    Statistical Analysis

    All analyses were conducted using R 3.4.4 (R Project for Statistical Computing), with 2-sided α = .05 as the threshold for statistical significance. Using multivariable-adjusted Cox proportional hazards models, this study calculated the hazard ratios (HRs) for mortality associated with prediagnostic frailty (at the 3-year WHI visit, before cancer diagnosis) and prediagnostic changes in frailty (from enrollment to the 3-year WHI visit). Analyses included women with cancer followed up from the time of diagnosis (time zero, as shown in the eFigure in the Supplement) to death or censoring (loss to follow-up or March 2018). Models for mortality after cancer diagnosis were stratified by age at enrollment (first exposure assessment) and adjusted for the covariates above, as well as for continuous age, cancer stage at diagnosis (local, regional, or distant), and year of diagnosis. When examining specific cancer sites, we adjusted for prognostic indicators specific to that cancer site. These included hormone receptor and ERBB2 (formerly HER2) status for breast cancer, non–small cell vs small cell for lung cancer, serous vs nonserous for ovarian cancer, colon vs rectum for colorectal cancer, and type 1 endometrioid adenocarcinomas vs type 2 uterine serous and clear cell carcinomas for endometrial cancer. We repeated these analyses with 3-year prediagnostic change in frailty status from enrollment to the 3-year WHI visit as the main exposure. As stated previously, all analyses excluded women who developed cancer before the 3-year WHI visit.

    In exploratory analyses examining whether women who develop cancer become increasingly frail, we fit linear mixed-effects models with frailty scores as a function of time since cancer diagnosis to calculate the time slope of change in the frailty score from enrollment through all subsequent measurements available as part of the WHI extension studies. These models included a random intercept and random slope for the change in frailty score as a function of time since diagnosis. We also included a time-varying indicator for date of first cancer diagnosis, at which time slopes could change. We then tested whether the rate of change in frailty (time slope and 95% CI) was altered from before to after cancer diagnosis using a likelihood ratio test to examine whether the addition of a linear spline with a knot at date of cancer diagnosis fit the data better than a more parsimonious model without a slope change. Subsequently, we evaluated whether the change in slope remained statistically significant after controlling for covariates (race/ethnicity, body mass index, pack-years of smoking, educational attainment, Charlson Comorbidity Index score, and any family history of cancer) and age and cancer stage at diagnosis. The physical function component of the SF-36 was most frequently updated; therefore, longitudinal assessments can be interpreted as the rate of change per year in physical function from before to after cancer diagnosis.

    For some women, frailty metrics at the 3-year WHI visit were far from cancer diagnosis. The median time from the 3-year WHI visit to cancer diagnosis was 7.4 years (range, 1 day to 19.8 years). To address this gap, we fit Cox proportional hazards models with time-updated frailty scores as the exposure using all available measurements to assess whether frailty closer to diagnosis was more strongly associated with survival (the median time from diagnosis to the closest prediagnostic frailty assessment was 1.6 years [range, 1 day to 17.3 years]). Finally, we considered whether our results were consistent across strata defined by age and cancer stage at diagnosis (local, regional, or distant) or excluding participants with comorbid conditions associated with frailty that were prevalent at the 3-year WHI visit (history of rheumatoid arthritis, history of congestive heart failure, emphysema, COPD, liver disease, and chronic kidney disease).

    Results

    This study included 7257 women in the WHI cohort who completed frailty assessments at enrollment and the 3-year WHI visit before cancer diagnosis and subsequently developed cancer. Cancer cases included 2644 breast cancers (36%), 822 lung cancers (11%), 691 colorectal cancers (10%), 445 endometrial cancers (6%), and 286 ovarian cancers (4%). At the time of the 3-year WHI visit (prediagnostic) frailty assessment, the mean (SD) age was 63 (7) years, and 44% (3205 of 7257) of women were 65 years or older. Sixteen percent (1161 of 7257) of participating women with cancer met the criteria for frailty, 29% (2129 of 7257) were prefrail, and 55% (3967 of 7257) were nonfrail. Women who were frail (vs nonfrail) at the 3-year WHI visit were older, more likely to be obese, and less likely to have a college education or a family income of $75 000 or more per year (eTable in the Supplement).

    Over a median follow-up 5.8 years (range, 1 day to 19.9 years) from cancer diagnosis, 3056 women (42%) died. Most deaths were from cancer (1985 [65%]), although the second most common cause of death was cardiovascular disease (314 [10%]). The survival probability after cancer diagnosis was lowest among frail women, intermediate among prefrail women, and highest among women who were nonfrail at enrollment (Figure 1). After multivariable adjustment, women who were frail (vs nonfrail) before cancer diagnosis had an increased risk of mortality after cancer diagnosis (HR, 1.40; 95% CI, 1.26-1.55; P for trend <.001) (Table 1).

    Over the 3 years after WHI enrollment, 21% (1537 of 7257) of women had sustained frailty (categorized as frail at both times), 22% (1578 of 7257) of women had worsening frailty (eg, from nonfrail to prefrail or frail and from prefrail to frail), and 12% (876 of 7257) had improved frailty (eg, from frail to prefrail or nonfrail and from prefrail to nonfrail). Women with sustained frailty or worsening frailty vs those who were consistently nonfrail (45% [3266 of 7257]) before cancer diagnosis had an increased risk of mortality after cancer diagnosis (HR, 1.25; 95% CI, 1.14-1.38 and 1.22; 95% CI, 1.11-1.34, respectively; P for trend <.001) (Table 2). Furthermore, cancer diagnosis served as an inflection point, after which women became increasingly frail. The rate of increase in physical frailty over time was statistically significantly higher after cancer diagnosis (Figure 2).

    When examining cancers separately, there was little evidence that the patterns of association differed by cancer site. Once diagnosed, frail women had a higher risk of mortality than nonfrail women for every cancer site. Similar results were observed for change in frailty status (Table 2).

    When we used time-updated assessments reflective of physical function closer to cancer diagnosis (Figure 3), results were consistent, and, in some cases, the risk of mortality associated with frailty was even greater. For example, in time-updated models, frail (vs nonfrail) women had a 71% increased risk of mortality after cancer diagnosis (HR, 1.71; 95% CI, 1.55-1.89).

    Frailty was consistently associated with an increased risk of mortality across subgroups defined by age at cancer diagnosis. The HRs of mortality for frail (vs nonfrail) women were 1.66 (95% CI, 1.16-2.38) for women younger than 65 years (n = 876), 1.43 (95% CI, 1.21-1.70) for women aged 65 to 74 years (n = 3013), and 1.37 (95% CI, 1.20-1.56) for women 75 years or older (n = 3368). Analyses stratified by cancer stage at diagnosis also yielded similar associations for local (n = 3856), regional (n = 1540), and distant (n = 1861) cancer stage at diagnosis. Restricting analyses to participants without history of rheumatoid arthritis, history of congestive heart failure, emphysema, COPD, liver disease, or chronic kidney disease at the 3-year WHI visit (n = 3445) did not materially change our results.

    Discussion

    In this study of 7257 postmenopausal women who developed cancer, prediagnostic frailty was associated with a 40% (HR, 1.40; 95% CI, 1.26-1.55) increased risk of mortality after cancer diagnosis (after adjusting for potential confounders). Similarly, sustained frailty (HR, 1.25; 95% CI, 1.14-1.38) or worsening frailty (HR, 1.22; 95% CI, 1.11-1.34) before cancer diagnosis was associated with a more than 20% increased risk of mortality after cancer diagnosis. Cancer diagnosis served as an inflection point, after which the rate of increase in frailty over time accelerated, reflecting a decline in physical function. Our study reinforces the importance of assessing and addressing frailty in patients with cancer.

    Prior studies3,12 have found that frailty is highly prevalent among patients with cancer and, as in the general population, is associated with increased treatment complications and postoperative and 1-year or 5-year mortality. Although heterogeneity in frailty definitions and outcomes makes direct comparison with prior studies challenging, our findings support those earlier results by documenting an increased long-term risk of mortality among patients with cancer who are frail and, for the first time to our knowledge, examining change in frailty status. Previous studies in patients with cancer have used other operational definitions of frailty, such as the cumulative deficit model24 or comprehensive geriatric assessments,25 both of which consider a broader range of functional ability and physical and cognitive health. We chose to evaluate a score conceptually aligned with the Fried criteria for frailty, which predominantly assesses physical frailty (our interest in this study) and has been shown to be sensitive to change after interventions in frail patients.26 Furthermore, the Fried frailty phenotype has been widely adopted and adapted because of its good construct validity6 and agreement with other approaches for assessing frailty.27

    To our knowledge, this study is the first to report that worsening prediagnostic frailty is detrimental to cancer survival. In addition, we found that the rate of increase in frailty (reflecting functional decline) accelerated after cancer diagnosis. Cancer and its treatments are hypothesized to accelerate aging, associated with increases in inflammation, cellular senescence, and other biological hallmarks of aging that underlie physical frailty.5,7,13-15 We found support for this hypothesis in that the rate of increase in functional decline over time accelerated after cancer diagnosis. This finding is consistent with prior literature demonstrating that functional status and overall health decline steeply after cancer diagnosis compared with age-matched controls without cancer.28,29 Although examining cancer treatments was outside the scope of the present analysis, much of this accelerated aging has been hypothesized to be associated with chemotherapy and radiotherapy, which shorten telomere length and increase p16 expression, proinflammatory cytokines, and resting energy expenditure.30

    In addition to screening for frailty in oncology settings, exercise and physical activity interventions may be beneficial in this setting. Frailty is modifiable, as evidenced by the fact that in this study a substantial proportion of women changed frailty categories over a 3-year period. In the general population, physical activity interventions (eg, resistance and aerobic exercise, brisk walking, or balance training) help to mitigate physical frailty and maintain independence in older adults.31 Nutrition support may also play a role, although results of multimodal interventions combining nutrition and physical activity have been mixed.32 Among cancer survivors, physical activity interventions are effective in preserving or increasing physical function,33 and findings of observational studies34-39 suggest that exercise may reduce mortality. Such supportive interventions might be considered if a patient is screened as frail at cancer diagnosis.

    Strengths and Limitations

    This study has notable strengths. It is one of the largest studies to date to examine phenotypic frailty and long-term cancer mortality. In addition, to our knowledge, it is one of the only studies to examine changes in functional decline over time.

    This study also has some limitations. First, frailty assessments were not timed to correspond to cancer diagnosis; therefore, the median time from the 3-year WHI visit frailty assessment to cancer diagnosis was 7.4 years. However, in analyses leveraging repeated frailty assessments collected over subsequent decades, the results were not only consistent but also indicated a stronger increased risk of mortality associated with prediagnostic frailty; this finding suggests that, while even measurements of frailty collected long before cancer diagnosis contain useful information, the closer in time to cancer diagnosis that frailty is assessed the stronger its association with cancer outcomes. Second, residual confounding is a possibility in all observational studies. Although we examined associations separately by cancer site and adjusted for age and stage at cancer diagnosis, preclinical cancer (ie, undiagnosed cancer that has not yet been detected clinically) could have contributed to frailty onset (eg, via unintentional weight loss). Although older age is associated with both frailty and mortality risk, we controlled for age at frailty assessment and age at cancer diagnosis in analyses and found similar associations across subgroups defined by age at cancer diagnosis. Data collection on specific cancer treatment modalities is ongoing. Further investigation on whether the observed accelerated functional decline after cancer diagnosis is because of cancer or its treatments is a critical area for future research.

    Conclusions

    This study found that sustained and worsening frailty before cancer diagnosis was associated with an increased risk of mortality after cancer diagnosis. Furthermore, the rate of decline in physical function accelerated after cancer diagnosis. These data support the importance of frailty assessment in patients with cancer to reduce and preempt worsening of physical frailty after cancer diagnosis.

    Back to top
    Article Information

    Accepted for Publication: July 1, 2020.

    Published: September 14, 2020. doi:10.1001/jamanetworkopen.2020.16747

    Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2020 Cespedes Feliciano EM et al. JAMA Network Open.

    Corresponding Author: Elizabeth M. Cespedes Feliciano, ScD, Division of Research, Kaiser Permanente Northern California, 2000 Broadway, Oakland, CA 94612 (elizabeth.m.cespedes@kp.org).

    Author Contributions: Mr Hohensee had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

    Concept and design: Cespedes Feliciano, Rosko.

    Acquisition, analysis, or interpretation of data: All authors.

    Drafting of the manuscript: Cespedes Feliciano.

    Critical revision of the manuscript for important intellectual content: All authors.

    Statistical analysis: Cespedes Feliciano, Hohensee.

    Obtained funding: Anderson, Paskett, Caan.

    Administrative, technical, or material support: Anderson, Paskett.

    Supervision: Cespedes Feliciano, Wallace.

    Conflict of Interest Disclosures: Dr Cespedes Feliciano reported receiving grants from the National Cancer Institute (NCI). Dr Rosko reported receiving grants and mentorship from the NCI; receiving travel support from VYXEOS; receiving personal fees from the Association of Community Cancer Centers, Medscape, and OncLive; and receiving other from Janssen, Millennium, Regeneron, and Acetylon. Dr Anderson reported receiving grants from the NCI and the National Heart, Lung, and Blood Institute (NHLBI). Dr Paskett reported receiving grants from the Breast Cancer Research Foundation, the Merck Foundation, and Pfizer. Dr Caan reported receiving grants from the NCI. No other disclosures were reported.

    Funding/Support: Dr Cespedes Feliciano was supported by NCI grants K01CA226155, R01CA240394, and R01CA251589. The WHI is supported by NCI grant UM1CA173642. The WHI is also supported by the NHLBI, NIH, and US Department of Health and Human Services through contracts N01WH22110, 24152, 32100-2, 32105-6, 32108-9, 32111-13, 32115, 32118-32119, 32122, 42107-26, 42129-32, and 44221.

    Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    References
    1.
    Bluethmann  SM, Mariotto  AB, Rowland  JH.  Anticipating the “silver tsunami”: prevalence trajectories and comorbidity burden among older cancer survivors in the United States.   Cancer Epidemiol Biomarkers Prev. 2016;25(7):1029-1036. doi:10.1158/1055-9965.EPI-16-0133 PubMedGoogle Scholar
    2.
    Clegg  A, Young  J, Iliffe  S, Rikkert  MO, Rockwood  K.  Frailty in elderly people.   Lancet. 2013;381(9868):752-762. doi:10.1016/S0140-6736(12)62167-9 PubMedGoogle Scholar
    3.
    Ethun  CG, Bilen  MA, Jani  AB, Maithel  SK, Ogan  K, Master  VA.  Frailty and cancer: implications for oncology surgery, medical oncology, and radiation oncology.   CA Cancer J Clin. 2017;67(5):362-377. doi:10.3322/caac.21406 PubMedGoogle Scholar
    4.
    Wildiers  H, Heeren  P, Puts  M,  et al.  International Society of Geriatric Oncology consensus on geriatric assessment in older patients with cancer.   J Clin Oncol. 2014;32(24):2595-2603. doi:10.1200/JCO.2013.54.8347 PubMedGoogle Scholar
    5.
    Woods  NF, LaCroix  AZ, Gray  SL,  et al; Women’s Health Initiative.  Frailty: emergence and consequences in women aged 65 and older in the Women’s Health Initiative Observational Study.   J Am Geriatr Soc. 2005;53(8):1321-1330. doi:10.1111/j.1532-5415.2005.53405.x PubMedGoogle Scholar
    6.
    Fried  LP, Tangen  CM, Walston  J,  et al; Cardiovascular Health Study Collaborative Research Group.  Frailty in older adults: evidence for a phenotype.   J Gerontol A Biol Sci Med Sci. 2001;56(3):M146-M156. doi:10.1093/gerona/56.3.M146 PubMedGoogle Scholar
    7.
     Erratum.   J Am Geriatr Soc. 2017;65(7):1631-1632. doi:10.1111/jgs.15006 PubMedGoogle Scholar
    8.
    Brundle  C, Heaven  A, Brown  L,  et al.  Convergent validity of the electronic frailty index.   Age Ageing. 2019;48(1):152-156. doi:10.1093/ageing/afy162 PubMedGoogle Scholar
    9.
    Clegg  A, Bates  C, Young  J,  et al.  Development and validation of an electronic frailty index using routine primary care electronic health record data.   Age Ageing. 2016;45(3):353-360. doi:10.1093/ageing/afw039 PubMedGoogle Scholar
    10.
    Rockwood  K, Mitnitski  A.  Frailty, fitness, and the mathematics of deficit accumulation.   Rev Clin Gerontol. 2007;17(1):1-12. doi:10.1017/S0959259807002353 Google Scholar
    11.
    Gilbert  T, Neuburger  J, Kraindler  J,  et al.  Development and validation of a Hospital Frailty Risk Score focusing on older people in acute care settings using electronic hospital records: an observational study.   Lancet. 2018;391(10132):1775-1782. doi:10.1016/S0140-6736(18)30668-8 PubMedGoogle Scholar
    12.
    Handforth  C, Clegg  A, Young  C,  et al.  The prevalence and outcomes of frailty in older cancer patients: a systematic review.   Ann Oncol. 2015;26(6):1091-1101. doi:10.1093/annonc/mdu540 PubMedGoogle Scholar
    13.
    Balducci  L, Fossa  SD.  Rehabilitation of older cancer patients.   Acta Oncol. 2013;52(2):233-238. doi:10.3109/0284186X.2012.744142 PubMedGoogle Scholar
    14.
    Mohile  SG, Xian  Y, Dale  W,  et al.  Association of a cancer diagnosis with vulnerability and frailty in older Medicare beneficiaries.   J Natl Cancer Inst. 2009;101(17):1206-1215. doi:10.1093/jnci/djp239 PubMedGoogle Scholar
    15.
    Pérez-Zepeda  MU, Cárdenas-Cárdenas  E, Cesari  M, Navarrete-Reyes  AP, Gutiérrez-Robledo  LM.  Cancer and frailty in older adults: a nested case-control study of the Mexican Health and Aging Study.   J Cancer Surviv. 2016;10(4):736-742. doi:10.1007/s11764-016-0519-6 PubMedGoogle Scholar
    16.
    Women’s Health Initiative Study Group.  Design of the Women’s Health Initiative clinical trial and observational study.   Control Clin Trials. 1998;19(1):61-109. doi:10.1016/S0197-2456(97)00078-0 PubMedGoogle Scholar
    17.
    Anderson  GL, Manson  J, Wallace  R,  et al.  Implementation of the Women’s Health Initiative study design.   Ann Epidemiol. 2003;13(9 suppl):S5-S17. doi:10.1016/S1047-2797(03)00043-7 PubMedGoogle Scholar
    18.
    McTiernan  A, Rossouw  J, Manson  J,  et al  Informed consent in the Women’s Health Initiative clinical trial and observational study.   J Womens Health. 1995;4(5):519-529. doi:10.1089/jwh.1995.4.519 Google Scholar
    19.
    Hays  RD, Morales  LS.  The RAND-36 measure of health-related quality of life.   Ann Med. 2001;33(5):350-357. doi:10.3109/07853890109002089 PubMedGoogle Scholar
    20.
    Ware  JE  Jr.  SF-36 health survey update.   Spine (Phila Pa 1976). 2000;25(24):3130-3139. doi:10.1097/00007632-200012150-00008 PubMedGoogle Scholar
    21.
    Zaslavsky  O, Zelber-Sagi  S, Gray  SL,  et al.  Comparison of frailty phenotypes for prediction of mortality, incident falls, and hip fracture in older women.   J Am Geriatr Soc. 2016;64(9):1858-1862. doi:10.1111/jgs.14233 PubMedGoogle Scholar
    22.
    National Cancer Institute. Surveillance, Epidemiology, and End Results Program. Accessed February 17, 2019. https://seer.cancer.gov/
    23.
    Stampfer  MJ, Willett  WC, Speizer  FE,  et al.  Test of the National Death Index.   Am J Epidemiol. 1984;119(5):837-839. doi:10.1093/oxfordjournals.aje.a113804 PubMedGoogle Scholar
    24.
    Rockwood  K, Howlett  SE, MacKnight  C,  et al.  Prevalence, attributes, and outcomes of fitness and frailty in community-dwelling older adults: report from the Canadian Study of Health and Aging.   J Gerontol A Biol Sci Med Sci. 2004;59(12):1310-1317. doi:10.1093/gerona/59.12.1310 PubMedGoogle Scholar
    25.
    Parker  SG, McCue  P, Phelps  K,  et al.  What is Comprehensive Geriatric Assessment (CGA)? an umbrella review.   Age Ageing. 2018;47(1):149-155. doi:10.1093/ageing/afx166 PubMedGoogle Scholar
    26.
    Li  CM, Chen  CY, Li  CY, Wang  WD, Wu  SC.  The effectiveness of a comprehensive geriatric assessment intervention program for frailty in community-dwelling older people: a randomized, controlled trial.   Arch Gerontol Geriatr. 2010;50(suppl 1):S39-S42. doi:10.1016/S0167-4943(10)70011-X PubMedGoogle Scholar
    27.
    Rockwood  K, Andrew  M, Mitnitski  A.  A comparison of two approaches to measuring frailty in elderly people.   J Gerontol A Biol Sci Med Sci. 2007;62(7):738-743. doi:10.1093/gerona/62.7.738 PubMedGoogle Scholar
    28.
    Petrick  JL, Foraker  RE, Kucharska-Newton  AM,  et al.  Trajectory of overall health from self-report and factors contributing to health declines among cancer survivors.   Cancer Causes Control. 2014;25(9):1179-1186. doi:10.1007/s10552-014-0421-3 PubMedGoogle Scholar
    29.
    Petrick  JL, Reeve  BB, Kucharska-Newton  AM,  et al.  Functional status declines among cancer survivors: trajectory and contributing factors.   J Geriatr Oncol. 2014;5(4):359-367. doi:10.1016/j.jgo.2014.06.002 PubMedGoogle Scholar
    30.
    Hodes  RJ, Sierra  F, Austad  SN,  et al.  Disease drivers of aging.   Ann N Y Acad Sci. 2016;1386(1):45-68. doi:10.1111/nyas.13299 PubMedGoogle Scholar
    31.
    Kidd  T, Mold  F, Jones  C,  et al.  What are the most effective interventions to improve physical performance in pre-frail and frail adults? a systematic review of randomised control trials.   BMC Geriatr. 2019;19(1):184. doi:10.1186/s12877-019-1196-x PubMedGoogle Scholar
    32.
    Hernández Morante  JJ, Gómez Martínez  C, Morillas-Ruiz  JM.  Dietary factors associated with frailty in old adults: a review of nutritional interventions to prevent frailty development.   Nutrients. 2019;11(1):E102. doi:10.3390/nu11010102 PubMedGoogle Scholar
    33.
    Daum  CW, Cochrane  SK, Fitzgerald  JD, Johnson  L, Buford  TW.  Exercise interventions for preserving physical function among cancer survivors in middle to late life.   J Frailty Aging. 2016;5(4):214-224.PubMedGoogle Scholar
    34.
    Arem  H, Moore  SC, Park  Y,  et al.  Physical activity and cancer-specific mortality in the NIH-AARP Diet and Health Study cohort.   Int J Cancer. 2014;135(2):423-431. doi:10.1002/ijc.28659 PubMedGoogle Scholar
    35.
    Borch  KB, Braaten  T, Lund  E, Weiderpass  E.  Physical activity before and after breast cancer diagnosis and survival: the Norwegian Women and Cancer cohort study.   BMC Cancer. 2015;15:967. doi:10.1186/s12885-015-1971-9 PubMedGoogle Scholar
    36.
    Demark-Wahnefried  W.  Cancer survival: time to get moving? data accumulate suggesting a link between physical activity and cancer survival.   J Clin Oncol. 2006;24(22):3517-3518. doi:10.1200/JCO.2006.06.6548 PubMedGoogle Scholar
    37.
    O’Donovan  G, Lee  IM, Hamer  M, Stamatakis  E.  Association of “weekend warrior” and other leisure time physical activity patterns with risks for all-cause, cardiovascular disease, and cancer mortality.   JAMA Intern Med. 2017;177(3):335-342. doi:10.1001/jamainternmed.2016.8014 PubMedGoogle Scholar
    38.
    Parekh  N, Lin  Y, Craft  LL, Vadiveloo  M, Lu-Yao  GL.  Longitudinal associations of leisure-time physical activity and cancer mortality in the Third National Health and Nutrition Examination Survey (1986-2006).   J Obes. 2012;2012:518358. doi:10.1155/2012/518358 PubMedGoogle Scholar
    39.
    Vainshelboim  B, Müller  J, Lima  RM,  et al.  Cardiorespiratory fitness, physical activity and cancer mortality in men.   Prev Med. 2017;100:89-94. doi:10.1016/j.ypmed.2017.04.014 PubMedGoogle Scholar
    ×