Background:

There is limited evidence describing associations between cancer and function in diverse cancer types and its relationship with mortality. We investigated the relationship between cancer and poor ambulatory function and associations between ambulatory function and subsequent mortality.

Methods:

Participants included 233,135 adults (n = 30,403 cancer and n = 202,732 cancer free) in the NIH-American Association of Retired Persons Diet and Health Study (1994–1996) who self-reported ambulatory function (e.g., walking pace and mobility disability: being unable to walk or walking at the slowest pace) in 2004–2006. Participants were followed for mortality from the assessment of ambulatory function through 2011. Multinomial logistic regression quantified the association between cancer and ambulatory function. We then explored the independent effects of walking pace and mobility disability in cancer survivors, and the joint effects of both a cancer diagnosis and poor ambulatory function on mortality using Cox proportional hazards models. Models explored type-specific associations across 15 cancer types.

Results:

Survivors had 42% greater odds of walking at the slowest pace [OR, 1.42 (confidence interval (CI), 1.30–1.54)] and 24% greater odds of mobility disability [OR, 1.24 (CI, 1.17–1.31)], compared with cancer-free participants, adjusting for baseline demographics, health indicators, and cancer type. Survivors reporting the slowest pace were at increased hazards than those who walked the fastest: all-cause mortality [HR, 2.22 (CI, 2.06–2.39)] and cancer mortality [HR, 2.12 (CI, 1.83–2.45)]. Similar trends emerged for mobility disability (HRs > 1.64). All-cause mortality associations were significant for more than nine cancer types.

Conclusions:

A diagnosis of cancer is associated with poorer ambulatory function, which is subsequently associated with increased mortality.

Impact:

Widespread efforts should target ambulatory function during cancer survivorship for survival benefits.

Most individuals diagnosed with cancer are over the age of 50 years (1) and face a unique burden of both aging- and cancer-related symptoms, such as impaired physical function, worsened quality of life, and premature mortality (2). Ambulatory function, as indicated by walking pace and major mobility disability, is critical for independence into old age (3) and often modifiable through behavior change, such as increased physical activity (4). Functional impairments after diagnosis and treatment for cancer have been documented (5–9), but much of this work has focused on a limited number of cancer types (10–13), less than 2 years of follow-up since diagnosis (5, 6, 11–13), and used objective measures of function (9). More recently, Bluethmann and colleagues (9) noted that survivors of cancer are more likely to walk at slower paces and show signs of disability. However, critical gaps remain surrounding the cancer–function association, specifically the reliance on objective measures of function. Such assessments are robust and important for their contribution to our understanding of the cancer–function relationship; however, they are often costly and limiting to many. There is a need to develop broad surveillance tools during cancer survivorship to identify those most in need of functional intervention. Furthermore, it is no longer only the “big four” cancer survivors who are living longer (14), yet there remains a paucity of evidence for differences in ambulatory function between the general population and survivors of rarer cancer types, as well as how long these differences may persist beyond the acute phase of diagnosis and treatment.

Poor ambulatory function has been consistently associated with worsened survival in healthy older adults (15–17). A pooled analysis from nine large cohorts identified increased mortality with slower walking pace in adults more than 65 years of age (16). Ambulatory function has often been referred to as the “sixth vital sign” given its validated and prognostic value for its functional perspective on health status (18). Such function–survival associations appear to extend to cancer survivors, as some work has identified increased risk of mortality with slower walking speeds (19–23). However, little has been done to investigate mortality risk by ambulatory function for a variety of cancer types, an important consideration in directing more precise interventions to promote functional health and longevity during survivorship (24).

To address these evidence gaps across 15 different cancer types, we examined: (i) whether a diagnosis of cancer is associated with lower ambulatory function and (ii) the independent effects of these functional domains on mortality in cancer survivors specifically, as well as the joint effects of both a cancer diagnosis and poor ambulatory function on mortality. We hypothesized that cancer survivors would be at increased risk of lower ambulatory function than cancer-free controls and that poor ambulatory function would be associated with increased risk of all-cause and cancer-specific mortality in cancer survivors.

Participants and study design

The NIH-American Association of Retired Persons (AARP) Diet and Health Study is a large prospective study of more than 500,000 AARP members between the ages of 50 and 71 years who responded to a baseline questionnaire assessing diet, medical history, and demographics in 1995–1996 (25). In 2004–2006, a follow-up questionnaire was sent to all surviving members of the cohort that assessed health status, lifestyle behaviors, body size, and many other factors, including ambulatory function (response rate, 44%). Figure 1 describes the basic study design, including key data collection points and the mortality follow-up period. The NIH-AARP Diet and Healthy Study was approved by the NCI (Bethesda, MD) Special Studies Institutional Review Board.

Figure 1.

Timeline. Timeline of data collection in the NIH-AARP Diet and Health Study.

Figure 1.

Timeline. Timeline of data collection in the NIH-AARP Diet and Health Study.

Close modal

Identification of cancer survivors

Cancer survivors were identified through linkage of the cohort to state cancer registries through December 31, 2006. Survivors were those with a registry-confirmed diagnosis (except nonmelanoma skin cancer) prior to completion of the follow-up questionnaire. Registries provided information on cancer diagnosis date, histology, stage, grade, and first course of treatment reported within 1 year of diagnosis. Assessment of cancer cases through the state registries has been estimated to be more than 90% (26).

Development of the analytic sample

Of the 566,398 respondents to the baseline questionnaire, 294,870 returned the follow-up questionnaire. To maintain temporality, we excluded cancer survivors diagnosed after the follow-up questionnaire (n = 35,721). To reduce the potential for confounding due to poor health status (27) or reporting inaccuracies, we further excluded survivors with metastatic disease (n = 1,944), recurrent disease (n = 138), nonregistry-confirmed diagnosis (n = 59), and those with more than an year between diagnosis and treatment (n = 1,476). All eligible participants who had missing ambulatory function information on the follow-up questionnaire were also excluded (n = 22,397), leaving a final analytic sample of 30,403 cancer survivors and 202,732 individuals free of cancer (233,135 total participants). Cancer type–specific analyses included only those cancer types with at least 100 cases.

Covariate assessment

The baseline questionnaire assessed age, sex, race, body mass index (BMI), highest level of educational attainment, and health status. Leisure-time physical activity and morbid conditions were self-reported on the follow-up questionnaire. A frailty index was created using the deficit accumulation approach (28), such that 26 morbid conditions reported at follow-up (see Supplementary Table S1) were summed, and then divided by the total number of conditions available. Individuals were then categorized as robust (<0.15), prefrail (0.15–0.24), mildly frail (0.25–0.34), or moderately-to-severely frail (≥0.35; ref. 29). Cancer type and treatment were assessed through state cancer registries and represent the first primary cancer type and first course of treatment (surgery, radiation, and chemotherapy).

Ambulatory function assessment

At follow-up, participants self-reported their normal walking pace [“easy” (<2 mph); “normal, average” (2–2.9 mph); “brisk” (3–3.9 mph); “very brisk, striding” (≥4 mph); and “unable to walk”]. Ambulatory function was evaluated in terms of walking pace and mobility disability. “Brisk” and “very brisk, striding” categories were collapsed together to better represent previous studies of gait speed in older adults (16). To prevent overestimation on analyses specific to walking pace, those who responded that they were “unable to walk” were removed (n = 6,501; “easy”, “normal”, and “≥brisk”). Mobility disability was classified as a report of being “unable to walk” or “easy” as has been done previously in the NIH-AARP Diet and Health Study (30). For clarity, an “easy” pace will be referred to as “slow” throughout the rest of the article, given it was the slowest pace reported by participants.

Mortality assessment

All-cause and cancer-specific mortality were assessed through linkage to the Social Security Administration Death Master File, as well as the National Death Index through December 31, 2011. International Classification of Diseases (ICD)-9 and ICD-10 codes were used to classify deaths due to all causes and with a cancer as the underlying cause. Ascertainment of deaths in this cohort is estimated to be more than 93% (31).

Statistical analysis

Because a large proportion of the original baseline cohort was not examined because of nonresponse on the follow-up questionnaire or participants moving out of areas covered by study-specific cancer registries, we used inverse probability weighting to account for the potential impact of differences between participants in the cohort at baseline and those who completed the follow-up questionnaire (32). Briefly, among all participants alive at the time of the follow-up questionnaire, we calculated a propensity score of completing the follow-up questionnaire using multiple logistic regression that included baseline age, sex, race, education, marital status, self-reported health status, smoking history, BMI, cancer, diabetes, heart disease, end-stage renal disease, gallbladder stones or disease, emphysema, osteoporosis, bone fracture after age 45 years, and polyps of colon or rectum. We then categorized participants into deciles on the basis of this propensity score and assigned individual weights that were inversely proportional to the number of individuals in their category that completed the questionnaire (33). These weights were then applied to all models.

To describe the association between a diagnosis of cancer and ambulatory function, multinomial logistic regression models estimated the odds of different walking paces for those with and without a cancer diagnosis. The odds of having mobility disability at follow-up by cancer status were assessed using logistic regression models. We further examined the associations between diagnosis for specific cancer types and ambulatory function. Models included those diagnosed with the cancer of interest and all controls in the cohort. Sensitivity analyses quantified how time since diagnosis may have contributed to the overall association using indicator variables (e.g., <1, 1–4.9, and ≥5 years). Models were adjusted for age, sex, race, BMI, self-reported health status, time under observation, frailty index, and cancer type (except in cancer-specific analyses).

To quantify independent effects of ambulatory function on mortality in cancer survivors, Cox proportional hazards regression models estimated adjusted HRs and 95% confidence intervals (CIs). Time from the follow-up questionnaire was the underlying time metric. We further examined these associations in specific cancer types; thus, we included only those diagnosed with the cancer of interest. Models were adjusted for all previous covariates plus leisure-time physical activity, time since diagnosis, and cancer treatment. Sensitivity analyses explored these associations by cancer treatment receipt (yes/no). To determine the joint effects of a cancer diagnosis and poor ambulatory function on mortality risk, we created indicator variables to directly compare the coexistence of both conditions against controls who walked the fastest and did not have mobility disability (reference groups). All statistical tests were conducted in SAS 9.4 (SAS Institute), and significance level was set to 0.05.

Participant characteristics are detailed in Table 1. Briefly, individuals were just over 60 years of age (Mage = 61.8 ± 5.36) at baseline, White (92.4%), male (56.7%), and generally healthy (56.4% reporting ≥“very good” health). Cancer survivors were mostly early stage (67.4% stage I disease) and previously had surgery for their disease (59.3%). Cancer types with ≥100 cases included breast, colon, endocrine, endometrial, melanoma, non-Hodgkin lymphoma, oral, ovarian, prostate, rectum, respiratory system, stomach, soft tissue, urinary, and vulva (Table 2; International Classification of Disease for Oncology, Third Edition; ref. 34). The average time between cancer diagnosis and assessment of ambulatory function in survivors was 4.6 ± 2.8 years. Follow-up time between ambulatory function assessment and death or censoring averaged 6.3 ± 1.7 years for survivors and 6.7 ± 1.1 years for controls.

Table 1.

Participant demographics and cancer-specific characteristics at baseline.

Full sampleNo cancerCancer
(N = 233,135)(n = 202,732)(n = 30,403)
 M (SD) or % M (SD) or % M (SD) or % 
Age (years) 61.82 (5.36) 61.61 (5.39) 63.22 (5.00) 
BMI 26.93 (4.91) 26.91 (4.94) 27.02 (4.74) 
White 92.41% 92.23% 93.60% 
Male 56.69% 55.13% 67.08% 
Self-reported health status 
 Excellent 18.98% 19.29% 16.96% 
 Very good 37.46% 37.52% 37.10% 
 Good 33.20% 32.93% 35.00% 
 Fair 8.19% 8.11% 8.69% 
 Poor 0.87% 0.87% 0.87% 
Frailty index 
 Robust 58.31% 58.83% 54.85% 
 Prefrail 32.35% 32.00% 34.65% 
 Mildly frail 7.74% 7.60% 8.71% 
 Moderately to severely frail 1.19% 1.17% 1.27% 
Cancer stage 
 0 — — 11.06% 
 I — — 67.43% 
 II — — 7.51% 
 III — — 5.14% 
 Unknown — — 8.86% 
Cancer treatment 
 Chemotherapy — — 10.41% 
 Radiation — — 27.20% 
 Surgery — — 59.29% 
Time between diagnosis and ambulatory function assessment (y) — — 4.55 (2.83) 
Full sampleNo cancerCancer
(N = 233,135)(n = 202,732)(n = 30,403)
 M (SD) or % M (SD) or % M (SD) or % 
Age (years) 61.82 (5.36) 61.61 (5.39) 63.22 (5.00) 
BMI 26.93 (4.91) 26.91 (4.94) 27.02 (4.74) 
White 92.41% 92.23% 93.60% 
Male 56.69% 55.13% 67.08% 
Self-reported health status 
 Excellent 18.98% 19.29% 16.96% 
 Very good 37.46% 37.52% 37.10% 
 Good 33.20% 32.93% 35.00% 
 Fair 8.19% 8.11% 8.69% 
 Poor 0.87% 0.87% 0.87% 
Frailty index 
 Robust 58.31% 58.83% 54.85% 
 Prefrail 32.35% 32.00% 34.65% 
 Mildly frail 7.74% 7.60% 8.71% 
 Moderately to severely frail 1.19% 1.17% 1.27% 
Cancer stage 
 0 — — 11.06% 
 I — — 67.43% 
 II — — 7.51% 
 III — — 5.14% 
 Unknown — — 8.86% 
Cancer treatment 
 Chemotherapy — — 10.41% 
 Radiation — — 27.20% 
 Surgery — — 59.29% 
Time between diagnosis and ambulatory function assessment (y) — — 4.55 (2.83) 

Abbreviations: BMI, body mass index; M, mean; SD, standard deviation.

Table 2.

Multinomial and logistic regression odds of poor ambulatory function of cancer survivors compared with cancer-free controls.

CasesWalkORCasesMobility disability
CancerNo.pace(95% CI)No.OR (95% CI)
All cancersa 29,468 Slow 1.42d (1.30–1.54) 30,403 1.24c (1.17–1.31) 
  Normal 1.20d (1.11–1.29)   
  Brisk (ref) 1.00   
Respiratoryb 1,257 Slow 2.62d (2.24–3.05) 1,317 1.95d (1.77–2.14) 
  Normal 1.49d (1.29–1.73)   
  Brisk (ref) 1.00   
Oral 451 Slow 1.97d (1.56–2.48) 469 1.60d (1.37–1.88) 
  Normal 1.33c (1.08–1.64)   
  Brisk (ref) 1.00   
Soft tissue 107 Slow 1.92c (1.13–3.24) 116 1.46c (1.04–2.05) 
  Normal 1.68c (1.05–2.67)   
  Brisk (ref) 1.00   
Stomach 124 Slow 1.72c (1.06–2.78) 128 1.18 (0.87–1.60) 
  Normal 1.49 (0.95–2.33)   
  Brisk (ref) 1.00   
Ovarian 127 Slow 1.44 (0.90–2.31) 135 1.23 (0.91–1.66) 
  Normal 1.37 (0.88–2.12)   
  Brisk (ref) 1.00   
Urinary 2,483 Slow 1.42d (1.28–1.57) 2,574 1.27d (1.18–1.36) 
  Normal 1.18c (1.08–1.29)   
  Brisk (ref) 1.00   
Colon 2,026 Slow 1.37d (1.22–1.53) 2,098 1.21d (1.12–1.30) 
  Normal 1.17c (1.05–1.30)   
  Brisk (ref) 1.00   
Rectum 795 Slow 1.32c (1.11–1.57) 827 1.15c (1.02–1.30) 
  Normal 1.22c (1.04–1.42)   
  Brisk (ref) 1.00   
Endocrine 275 Slow 1.24 (0.92–1.66) 284 1.20 (0.97–1.48) 
  Normal 1.05 (0.81–1.38)   
  Brisk (ref) 1.00   
Endometrial 784 Slow 1.18 (0.96–1.43) 844 1.15c (1.02–1.31) 
  Normal 1.04 (0.86–1.25)   
  Brisk (ref) 1.00   
Non-Hodgkin lymphoma 635 Slow 1.12 (0.92–1.37) 659 1.11 (0.96–1.28) 
  Normal 1.04 (0.87–1.24)   
  Brisk (ref) 1.00   
Breast 5,045 Slow 1.11c (1.03–1.20) 5,298 1.07c (1.02–1.12) 
  Normal 1.06 (0.99–1.14)   
  Brisk (ref) 1.00   
Prostate 11,801 Slow 1.05c (1.01–1.10) 11,997 1.05c (1.01–1.08) 
  Normal 1.00 (0.97–1.04)   
  Brisk (ref) 1.00   
Vulva 109 Slow 0.95 (0.59–1.53) 116 0.91 (0.67–1.25) 
  Normal 1.06 (0.68–1.63)   
  Brisk (ref) 1.00   
Melanoma 2,428 Slow 0.86c (0.77–0.95) 2,474 0.88c (0.81–0.95) 
  Normal 0.95 (0.88–1.04)   
  Brisk (ref) 1.00   
CasesWalkORCasesMobility disability
CancerNo.pace(95% CI)No.OR (95% CI)
All cancersa 29,468 Slow 1.42d (1.30–1.54) 30,403 1.24c (1.17–1.31) 
  Normal 1.20d (1.11–1.29)   
  Brisk (ref) 1.00   
Respiratoryb 1,257 Slow 2.62d (2.24–3.05) 1,317 1.95d (1.77–2.14) 
  Normal 1.49d (1.29–1.73)   
  Brisk (ref) 1.00   
Oral 451 Slow 1.97d (1.56–2.48) 469 1.60d (1.37–1.88) 
  Normal 1.33c (1.08–1.64)   
  Brisk (ref) 1.00   
Soft tissue 107 Slow 1.92c (1.13–3.24) 116 1.46c (1.04–2.05) 
  Normal 1.68c (1.05–2.67)   
  Brisk (ref) 1.00   
Stomach 124 Slow 1.72c (1.06–2.78) 128 1.18 (0.87–1.60) 
  Normal 1.49 (0.95–2.33)   
  Brisk (ref) 1.00   
Ovarian 127 Slow 1.44 (0.90–2.31) 135 1.23 (0.91–1.66) 
  Normal 1.37 (0.88–2.12)   
  Brisk (ref) 1.00   
Urinary 2,483 Slow 1.42d (1.28–1.57) 2,574 1.27d (1.18–1.36) 
  Normal 1.18c (1.08–1.29)   
  Brisk (ref) 1.00   
Colon 2,026 Slow 1.37d (1.22–1.53) 2,098 1.21d (1.12–1.30) 
  Normal 1.17c (1.05–1.30)   
  Brisk (ref) 1.00   
Rectum 795 Slow 1.32c (1.11–1.57) 827 1.15c (1.02–1.30) 
  Normal 1.22c (1.04–1.42)   
  Brisk (ref) 1.00   
Endocrine 275 Slow 1.24 (0.92–1.66) 284 1.20 (0.97–1.48) 
  Normal 1.05 (0.81–1.38)   
  Brisk (ref) 1.00   
Endometrial 784 Slow 1.18 (0.96–1.43) 844 1.15c (1.02–1.31) 
  Normal 1.04 (0.86–1.25)   
  Brisk (ref) 1.00   
Non-Hodgkin lymphoma 635 Slow 1.12 (0.92–1.37) 659 1.11 (0.96–1.28) 
  Normal 1.04 (0.87–1.24)   
  Brisk (ref) 1.00   
Breast 5,045 Slow 1.11c (1.03–1.20) 5,298 1.07c (1.02–1.12) 
  Normal 1.06 (0.99–1.14)   
  Brisk (ref) 1.00   
Prostate 11,801 Slow 1.05c (1.01–1.10) 11,997 1.05c (1.01–1.08) 
  Normal 1.00 (0.97–1.04)   
  Brisk (ref) 1.00   
Vulva 109 Slow 0.95 (0.59–1.53) 116 0.91 (0.67–1.25) 
  Normal 1.06 (0.68–1.63)   
  Brisk (ref) 1.00   
Melanoma 2,428 Slow 0.86c (0.77–0.95) 2,474 0.88c (0.81–0.95) 
  Normal 0.95 (0.88–1.04)   
  Brisk (ref) 1.00   

Note: Covariates: age, sex, race, highest level of education, years of follow-up, BMI, self-reported health status, and frailty index. Sorted by walking pace strength of association.

Abbreviations: BMI, body mass index; M, mean; SD, standard deviation.

aAlso adjusted for cancer type.

bRespiratory system includes cancers of the nose, nasal cavity, middle ear, larynx, lung and bronchus, pleura, trachea, mediastinum, and other respiratory.

cP = 0.05.

dP < 0.001.

Associations between cancer and ambulatory function

A diagnosis of cancer was associated with increased odds of poor ambulatory function (Table 2, sorted by walking pace strength of association). Relative to the fastest walking pace, individuals with a diagnosis of cancer had 42% higher odds of walking at the slowest pace compared with adults without a cancer diagnosis [OR, 1.42 (1.30–1.54)]. Similarly, survivors had significantly higher odds of mobility disability at follow-up [OR, 1.24 (1.17–1.31)], after adjustment for demographics, BMI, health status, and cancer type. Sensitivity analyses indicated that regardless of time since diagnosis, survivors had increased odds of poor ambulatory function compared with controls (P < 0.001; see Supplementary Table S2).

The association between cancer diagnosis and lower ambulatory function persisted in nine different cancer types compared with controls (Table 2). A diagnosis of breast, colon, oral, prostate, rectal, respiratory, soft tissue, stomach, and urinary cancer was significantly associated with higher odds of slow walking (ORs > 1.05). A similar pattern emerged for mobility disability, such that a diagnosis of breast, colon, endometrial, oral, prostate, rectal, respiratory, soft tissue, and urinary cancer was associated with greater risk of being disabled than controls. Across both ambulatory function indicators, the strongest associations were evident for respiratory and oral cancers (ORs > 1.60). Survivors with melanoma had significantly lower odds of poor ambulatory function than controls (ORs < 0.88).

Associations between ambulatory function and mortality in cancer survivors

We found significant independent effects of ambulatory function on mortality. Slower walking paces were significantly associated with higher risk of all-cause and cancer-specific mortality, adjusting for demographics and cancer characteristics (Fig. 2). Compared with the fastest walking survivors, those reporting slower walking had over a doubling of risk for all-cause mortality [HR, 2.22 (2.06–2.39)]. Mobility disability after cancer was also significantly associated with increased hazards of all-cause mortality [HR, 1.80 (1.73–1.88)]. This pattern of association persisted for cancer-specific mortality. Sensitivity analyses indicated no significant differences by treatment receipt (Pinteraction > 0.106; see Supplementary Table S3).

Figure 2.

Walking pace, mobility disability, and hazards of all-cause and cancer mortality in cancer survivors. Covariates: age, sex, race, highest level of education, years of follow-up, BMI, self-reported health status, cancer type, leisure time physical activity, cancer treatment, frailty index, and time between cancer diagnosis and follow-up questionnaire.

Figure 2.

Walking pace, mobility disability, and hazards of all-cause and cancer mortality in cancer survivors. Covariates: age, sex, race, highest level of education, years of follow-up, BMI, self-reported health status, cancer type, leisure time physical activity, cancer treatment, frailty index, and time between cancer diagnosis and follow-up questionnaire.

Close modal

Of the 15 cancer types examined, poor ambulatory function was associated with increased mortality risk across most cancers (Fig. 3; Supplementary Tables S4 and S5). In nine cancer types, survivors who reported the slowest walking pace were at significantly increased risk of all-cause mortality, including breast, colon, melanoma, non-Hodgkin lymphoma, oral, prostate, rectal, respiratory, and urinary cancers (HRs > 1.91). More robust associations were evidenced for mobility disability. Thirteen different cancers had significant associations between mobility disability and all-cause mortality, including breast, colon, endometrial, endocrine, melanoma, non-Hodgkin lymphoma, oral, ovarian, prostate, rectal, respiratory, stomach, and urinary cancers (HRs > 1.34).

Figure 3.

Hazards of all-cause mortality in cancer survivors: reporting a slow pace compared with those reporting ≥brisk walking pace (A) and reporting mobility disability compared with those without mobility disability (B). Covariates: age, sex, race, highest level of education, years of follow-up, BMI, self-reported health status, leisure time physical activity, cancer treatment, frailty index, and time between cancer diagnosis and follow-up questionnaire.

Figure 3.

Hazards of all-cause mortality in cancer survivors: reporting a slow pace compared with those reporting ≥brisk walking pace (A) and reporting mobility disability compared with those without mobility disability (B). Covariates: age, sex, race, highest level of education, years of follow-up, BMI, self-reported health status, leisure time physical activity, cancer treatment, frailty index, and time between cancer diagnosis and follow-up questionnaire.

Close modal

Joint associations of cancer and ambulatory function on mortality

Finally, we examined the joint effects of poor ambulatory function and cancer on mortality in both cancer survivors and controls and found a statistically significant interaction between ambulatory function and a previous cancer diagnosis (P < 0.001; Fig. 4). Lower ambulatory function was associated with increased risk of all-cause mortality in both survivors and controls; however, the adverse risk association was much stronger in survivors. For example, compared with the fastest walking controls, cancer survivors who walked the slowest had a much higher HR than controls who walked the slowest (10.37 vs. 3.91). Similar patterns emerged for mobility disability.

Figure 4.

Hazards of all-cause mortality by coexistence of poor ambulatory function and cancer: walking pace (A) and mobility disability (B).

Figure 4.

Hazards of all-cause mortality by coexistence of poor ambulatory function and cancer: walking pace (A) and mobility disability (B).

Close modal

Our findings suggest that diagnosis and treatment for cancer are associated with poor ambulatory function up to 5 years later. These results were consistent across nine different types of cancer, suggesting that more cancer types may predispose individuals to poor ambulatory function than studied previously. Lower ambulatory function placed cancer survivors at increased risk of death, highlighting the importance of these functional domains during survivorship. Importantly, our results also suggest that risks associated with lower ambulatory function may be two to three times stronger for early mortality in survivors compared with the general population.

Our finding that a diagnosis of cancer was associated with poor ambulatory function is consistent with the literature highlighting pervasive and detrimental effects of cancer on functional health (5–8, 21, 35, 36). To date, most of this work has been confined to common cancer types, such as breast cancer (8, 37). The novel contributions of this study allowed us to examine longitudinal associations between cancer and ambulatory function across 15 different cancer types. A wide variety of cancer types were associated with lower ambulatory function, including those of the gastrointestinal tract, hormonally related cancers, and even cancers with watch-and-wait treatment recommendations (prostate). Interestingly, melanoma was associated with lower odds of poor ambulatory function, which may be due to the known relationship between leisure-time physical activity and increased melanoma risk (38). The biological and/or psychosocial mechanisms of the differential odds of ambulatory function in specific cancer types should be explored further. This sample of cancer survivors is on average several years beyond treatment cessation, and a better understanding of both pre- and postdiagnosis behavioral, biological, and cancer-specific factors will be important to fully envision these complex relationships. Specifically, future research would do well to explore changes in ambulatory function across the cancer continuum, not just posttreatment, to better identify where functional changes occur for each cancer type and optimal time(s) for intervention.

Walking pace and mobility disability were also independently associated with increased mortality risk in cancer survivors, just as ambulatory function has emerged as a robust predictor of survival in healthy older adults (16, 39). These findings corroborate a recent analysis of self-reported mobility disability and mortality among survivors. (21) Brown and colleagues identified 215% and 249% increased risk of all-cause and cancer mortality with mobility disability (as assessed via one's perceived difficulty in walking for a quarter of a mile), respectively, among 1,458 cancer survivors in the National Health and Nutrition Examination Survey. Other studies have also noted increased mortality hazards in survivors with slower walking paces, albeit with binary definitions and limited cancer types (20, 22). Importantly, we explored the function–mortality relationship among multiple walking pace categories and 15 different cancer types, highlighting significant associations among nine cancers for walking pace and 13 cancers for mobility disability. Previously unexplored, these findings provide a foundation for future research into potential cancer-specific mechanisms that may be increasing mortality risk. Despite chemotherapy's notorious reputation for deleterious late effects, sensitivity analyses suggested that these results cannot be explained by treatment regimen alone; future research would do well to investigate other cancer type- or aging-specific factors that may explain the function–survival relationship. That said, poor function was associated with mortality in most cancer types studied herein, suggesting that universal efforts toward improving ambulatory function during survivorship are worthwhile. Such efforts may include physical activity, which has been consistently associated with improved gait speed during survivorship (40).

Indeed, our joint effects of cancer and function on all-cause mortality point to substantially increased risk of death in survivors with poor functioning compared with healthy controls. If replicated, these findings suggest that ambulatory function may be more important for survival in cancer survivors for reasons we do not yet fully understand. Potentially due to an accelerated aging trajectory (2), coupled with other cancer-specific characteristics unavailable in this sample (e.g., specific chemotherapies, multiple treatments, and medications), clarifying the underlying biological mechanisms of these findings is warranted.

Of further interest is the self-reported nature of ambulatory function in this analysis and its implications for widespread surveillance of cancer survivors. Historically, objective measures of walking pace (20, 22) and clinical indicators, such as the Eastern Cooperative Oncology Group (ECOG) performance status (41) and Karnofsky performance status (42), have served as prognostic factors of survival in cancer survivors. Despite their importance, implementing these measures in practice may be unrealistic for many older cancer survivors due to burden, cost, and time (21), and they certainly are not applicable to broader population assessment. While ECOG and Karnofsky performance statuses or other objective functional batteries were unavailable in this sample, our findings suggest that simple patient-reported outcomes, such as one question about walking pace, may be efficient alternatives for screening survivors on a large scale to identify those who may need follow-up with more rigorous assessments of ambulatory function. Future research should include multiple functional measures to determine the independent prognostic ability of self-reported walking pace and/or mobility disability for determining survival.

Our results should be interpreted in the context of the study's strengths and limitations. The primary strength of this analysis is its prospective nature, allowing us to test temporal associations between cancer and future ambulatory function, and ambulatory function and future mortality. Our large sample size also allowed us to explore these relationships in multiple cancer types that had previously been unexplored. We acknowledge that our analytic sample was comprised of less than the original NIH-AARP cohort, thus results may be more relevant to individuals who filled out the follow-up questionnaire. However, we conducted inverse probability weighting to statistically account for these differences. Furthermore, most survivors were early stage and multiple years postdiagnosis. Limited treatment information available from cancer registries also made it difficult to understand the specific regimens that may be detrimental to ambulatory function. It is nonetheless encouraging that such associations emerged in a relatively healthy sample, suggesting that the cancer–function relationship may be more pronounced, and potentially even more important, in more diverse samples. It is also worth noting that other unmeasured factors (e.g., mild cognitive impairment or dementia) may confound the relationships reported herein. However, we included a constellation of potential confounders to estimate the association to the best of our ability with the data available. Finally, despite research highlighting the robust association between self-reported and objective measures of ambulatory function (43), we are unable to say with certainty that self-reported walking pace in this study paralleled true gait speed; however, these subjective measures allowed for broad, cost-effective identification of at-risk survivors.

To our knowledge, this is the first prospective analysis to have considered both the role of a cancer diagnosis on ambulatory function and the implications of poor function on subsequent survival in 15 different cancer types. Our findings confirm an increased risk of poor ambulatory function after cancer, as well as survivors' increased risk of mortality with lower functional health. Notably, these associations persisted in at least nine different cancer types, highlighting the need for comprehensive surveillance and targeting of ambulatory function during survivorship, potentially through cancer rehabilitation and/or behavioral change (e.g., physical activity). Given that cancer survivors with poor ambulatory function had two to three times greater mortality risk than their cancer-free peers, public health efforts should focus on identifying and targeting survivors who are at-risk for poor functional health to ultimately improve survival.

No disclosures were reported.

The views expressed herein are solely those of the authors and do not necessarily reflect those of the FCDC or FDOH. The Pennsylvania Department of Health specifically disclaims responsibility for any analyses, interpretations or conclusions.

E.A. Salerno: Conceptualization, formal analysis, methodology, writing–original draft, writing–review and editing. P.F. Saint-Maurice: Formal analysis, writing–review and editing. E.A. Willis: Writing–review and editing. S.C. Moore: Writing–review and editing. L. DiPietro: Writing–review and editing. C.E. Matthews: Conceptualization, supervision, methodology, writing–review and editing.

This research was supported (in part) by the Intramural Research Program of the NIH, NCI. Cancer incidence data from the Atlanta metropolitan area were collected by the Georgia Center for Cancer Statistics, Department of Epidemiology, Rollins School of Public Health, Emory University (Atlanta, GA). Cancer incidence data from California were collected by the California Cancer Registry, California Department of Public Health's Cancer Surveillance and Research Branch (Sacramento, CA). Cancer incidence data from the Detroit metropolitan area were collected by the Michigan Cancer Surveillance Program, Community Health Administration (Lansing, MI). The Florida cancer incidence data used in this report were collected by the Florida Cancer Data System (Miami, FL) under contract with the Florida Department of Health (Tallahassee, FL). Cancer incidence data from Louisiana were collected by the Louisiana Tumor Registry, Louisiana State University Health Sciences Center School of Public Health (New Orleans, LA). Cancer incidence data from New Jersey were collected by the New Jersey State Cancer Registry, The Rutgers Cancer Institute of New Jersey (New Brunswick, NJ). Cancer incidence data from North Carolina were collected by the North Carolina Central Cancer Registry (Raleigh, NC). Cancer incidence data from Pennsylvania were supplied by the Division of Health Statistics and Research, Pennsylvania Department of Health (Harrisburg, PA). Cancer incidence data from Arizona were collected by the Arizona Cancer Registry, Division of Public Health Services, Arizona Department of Health Services (Phoenix, AZ). Cancer incidence data from Texas were collected by the Texas Cancer Registry, Cancer Epidemiology and Surveillance Branch, Texas Department of State Health Services (Austin, TX). Cancer incidence data from Nevada were collected by the Nevada Central Cancer Registry, Division of Public and Behavioral Health, State of Nevada Department of Health and Human Services (Carson City, NV). We are indebted to the participants in the NIH-AARP Diet and Health Study for their outstanding cooperation. We also thank Sigurd Hermansen and Kerry Grace Morrissey from Westat for study outcomes ascertainment and management and Leslie Carroll at Information Management Services for data support and analysis.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Siegel
RL
,
Miller
KD
,
Jemal
A
. 
Cancer statistics, 2019
.
CA Cancer J Clin
2019
;
69
:
7
34
.
2.
Henderson
TO
,
Ness
KK
,
Cohen
HJ
. 
Accelerated aging among cancer survivors: from pediatrics to geriatrics
.
Am Soc Clin Oncol Educ B
2014
;
34
:
e423
30
.
3.
Lowry
KA
,
Vallejo
AN
,
Studenski
SA
. 
Successful aging as a continuum of functional independence: lessons from physical disability models of aging
.
Aging Dis
2012
;
3
:
5
15
.
4.
Nelson
ME
,
Rejeski
WJ
,
Blair
SN
,
Duncan
PW
,
Judge
JO
,
King
AC
, et al
Physical activity and public health in older adults: recommendation from the American College of Sports Medicine and the American Heart Association
.
Circulation
2007
;
116
:
1094
105
.
5.
Avis
NE
,
Deimling
GT
. 
Cancer survivorship and aging
.
Cancer
2008
;
113
:
3519
29
.
6.
Bellizzi
KM
,
Mustian
KM
,
Palesh
OG
,
Diefenbach
M
. 
Cancer survivorship and aging
.
Cancer
2008
;
113
:
3530
9
.
7.
Demark-Wahnefried
W
,
Pinto
BM
,
Gritz
ER
. 
Promoting health and physical function among cancer survivors: potential for prevention and questions that remain
.
J Clin Oncol
2006
;
24
:
5125
31
.
8.
Petrick
JL
,
Reeve
BB
,
Kucharska-Newton
AM
,
Foraker
RE
,
Platz
EA
,
Stearns
SC
, et al
Functional status declines among cancer survivors: Trajectory and contributing factors
.
J Geriatr Oncol
2014
;
5
:
359
67
.
9.
Bluethmann
SM
,
Flores
E
,
Campbell
G
,
Klepin
HD
. 
Mobility device use and mobility disability in U.S. Medicare beneficiaries with and without cancer history
.
J Am Geriatr Soc
2020
;
68
:
2872
80
.
10.
Given
CW
,
Given
B
,
Azzouz
F
,
Stommel
M
,
Kozachik
S
. 
Comparison of changes in physical functioning of elderly patients with new diagnoses of cancer
.
Med Care
2000
;
38
:
482
93
.
11.
Ganz
PA
,
Guadagnoli
E
,
Landrum
MB
,
Lash
TL
,
Rakowski
W
,
Silliman
RA
. 
Breast cancer in older women: quality of life and psychosocial adjustment in the 15 months after diagnosis
.
J Clin Oncol
2003
;
21
:
4027
33
.
12.
Arndt
V
,
Merx
H
,
Stürmer
T
,
Stegmaier
C
,
Ziegler
H
,
Brenner
H
. 
Age-specific detriments to quality of life among breast cancer patients one year after diagnosis
.
Eur J Cancer
2004
;
40
:
673
80
.
13.
Kroenke
CH
,
Rosner
B
,
Chen
WY
,
Kawachi
I
,
Colditz
GA
,
Holmes
MD
. 
Functional impact of breast cancer by age at diagnosis
.
J Clin Oncol
2004
;
22
:
1849
56
.
14.
Miller
KD
,
Nogueira
L
,
Mariotto
AB
,
Rowland
JH
,
Yabroff
KR
,
Alfano
CM
, et al
Cancer treatment and survivorship statistics, 2019
.
CA Cancer J Clin
2019
;
69
:
363
85
.
15.
Hirvensalo
M
,
Rantanen
T
,
Heikkinen
E
. 
Mobility difficulties and physical activity as predictors of mortality and loss of independence in the community-living older population
.
J Am Geriatr Soc
2000
;
48
:
493
8
.
16.
Studenski
S
,
Perera
S
,
Patel
K
,
Faulkner
K
,
Inzitari
M
,
Brach
J
, et al
Gait speed and survival in older adults
.
JAMA
2011
;
305
:
50
.
17.
Stamatakis
E
,
Kelly
P
,
Strain
T
,
Murtagh
EM
,
Ding
D
,
Murphy
MH
. 
Self-rated walking pace and all-cause, cardiovascular disease and cancer mortality: individual participant pooled analysis of 50 225 walkers from 11 population British cohorts
.
Br J Sports Med
2018
;
52
:
761
8
.
18.
Fritz
S
,
Lusardi
MM
. 
White paper: walking speed: the sixth vital sign
.
J Geriatr Phys Ther
2009
;
32
:
2
5
.
19.
Brown
JC
,
Harhay
MO
,
Harhay
MN
. 
Physical function as a prognostic biomarker among cancer survivors
.
Br J Cancer
2015
;
112
:
194
8
.
20.
Pamoukdjian
F
,
Lévy
V
,
Sebbane
G
,
Boubaya
M
,
Landre
T
,
Bloch-Queyrat
C
, et al
Slow gait speed is an independent predictor of early death in older cancer outpatients: results from a prospective cohort study
.
J Nutr Health Aging
2017
;
21
:
202
6
.
21.
Brown
JC
,
Harhay
MO
,
Harhay
MN
. 
Self-reported major mobility disability and mortality among cancer survivors
.
J Geriatr Oncol
2018
;
9
:
459
63
.
22.
Cesari
M
,
Cerullo
F
,
Zamboni
V
,
Di Palma
R
,
Scambia
G
,
Balducci
L
, et al
Functional status and mortality in older women with gynecological cancer
.
J Gerontol A Biol Sci Med Sci
2013
;
68
:
1129
33
.
23.
Klepin
HD
,
Geiger
AM
,
Tooze
JA
,
Newman
AB
,
Colbert
LH
,
Bauer
DC
, et al
Physical performance and subsequent disability and survival in older adults with malignancy: results from the health, aging and body composition study
.
J Am Geriatr Soc
2010
;
58
:
76
82
.
24.
Siegel
RL
,
Miller
KD
,
Jemal
A
. 
Cancer statistics, 2020
.
CA Cancer J Clin
2020
;
70
:
7
30
.
25.
Schatzkin
A
,
Subar
AF
,
Thompson
FE
,
Harlan
LC
,
Tangrea
J
,
Hollenbeck
AR
, et al
Design and serendipity in establishing a large cohort with wide dietary intake distributions
.
Am J Epidemiol
2001
;
154
:
1119
25
.
26.
Michaud
D
,
Midthune
D
,
Hermansen
S
,
Leitzmman
MF
,
Harlan
LC
,
Kipnis
V
, et al
Comparison of cancer registry case ascertainment with SEER estimates and self-reporting in a subset of the NIH-AARP Diet and Health Study
.
J Registry Manag
2005
;
32
:
70
5
.
27.
Matthews
CE
,
Troiano
RP
,
Salerno
EA
,
Berrigan
D
,
Patel
SB
,
Shiroma
EJ
, et al
Exploration of confounding due to poor health in an accelerometer–mortality study
.
Med Sci Sports Exerc
2020
;
52
:
2546
53
.
28.
Theou
O
,
Walston
J
,
Rockwood
K
. 
Operationalizing frailty using the frailty phenotype and deficit accumulation approaches
. In:
interdisciplinary topics in gerontology and geriatrics
.
Basel, Switzerland
,
Karger Publishers
; 
2015
. p.
66
73
.
29.
Kim
DH
,
Glynn
RJ
,
Avorn
J
,
Lipsitz
LA
,
Rockwood
K
,
Pawar
A
, et al
Validation of a claims-based frailty index against physical performance and adverse health outcomes in the health and retirement study
.
J Gerontol A Biol Sci Med Sci
2019
;
74
:
1271
6
.
30.
DiPietro
L
,
Jin
Y
,
Talegawkar
S
,
Matthews
CE
. 
The joint associations of sedentary time and physical activity with mobility disability in older people: the NIH-AARP diet and health study
.
J Gerontol A Biol Sci Med Sci
2018
;
73
:
532
8
.
31.
Rich-Edwards
JW
,
Corsano
KA
,
Stampfer
MJ
. 
Test of the national death index and Equifax nationwide death search
.
Am J Epidemiol
1994
;
140
:
1016
9
.
32.
Seaman
SR
,
White
IR
. 
Review of inverse probability weighting for dealing with missing data
.
Stat Methods Med Res
2013
;
22
:
278
95
.
33.
Curtis
LH
,
Hammill
BG
,
Eisenstein
EL
,
Kramer
JM
,
Anstrom
KJ
. 
Using inverse probability-weighted estimators in comparative effectiveness analyses with observational databases
.
Med Care
2007
;
45
:
S103
7
.
34.
Fritz
AG
, editor.
International Classification of Disease for Oncology: ICD-O
.
Geneva, Switzerland
, 
2013
.
35.
Keating
NL
,
Nørredam
M
,
Landrum
MB
,
Huskamp
HA
,
Meara
E
. 
Physical and mental health status of older long-term cancer survivors
.
J Am Geriatr Soc
2005
;
53
:
2145
52
.
36.
Deimling
GT
,
Sterns
S
,
Bowman
KF
,
Kahana
B
. 
Functioning and activity participation restrictions among older adult, long-term cancer survivors
.
Cancer Invest
2007
;
25
:
106
16
.
37.
Sweeney
C
,
Schmitz
KH
,
Lazovich
D
,
Virnig
BA
,
Wallace
RB
,
Folsom
AR
. 
Functional limitations in elderly female cancer survivors
.
J Natl Cancer Inst
2006
;
98
:
521
9
.
38.
Moore
SC
,
Lee
I-M
,
Weiderpass
E
,
Campbell
PT
,
Sampson
JN
,
Kitahara
CM
, et al
Association of leisure-time physical activity with risk of 26 types of cancer in 1.44 million adults
.
JAMA Intern Med
2016
;
176
:
816
.
39.
Hardy
SE
,
Perera
S
,
Roumani
YF
,
Chandler
JM
,
Studenski
SA
. 
Improvement in usual gait speed predicts better survival in older adults
.
J Am Geriatr Soc
2007
;
55
:
1727
34
.
40.
Kirkham
AA
,
Bland
KA
,
Sayyari
S
,
Campbell
KL
,
Davis
MK
. 
Clinically relevant physical benefits of exercise interventions in breast cancer survivors
.
Curr Oncol Rep
2016
;
18
:
12
.
41.
Zubrod
CG
,
Schneiderman
M
,
Frei
E
,
Brindley
C
,
Lennard Gold
G
,
Shnider
B
, et al
Appraisal of methods for the study of chemotherapy of cancer in man: Comparative therapeutic trial of nitrogen mustard and triethylene thiophosphoramide
.
J Chronic Dis
1960
;
11
:
7
33
.
42.
Karnofsky
DA
. 
The clinical evaluation of chemotherapeutic agents in cancer. Evaluation of chemotherapeutic agents
.
Available from
: https://ci.nii.ac.jp/naid/10005058071.
43.
Brown
JC
,
Harhay
MO
,
Harhay
MN
. 
Patient-reported versus objectively-measured physical function and mortality risk among cancer survivors
.
J Geriatr Oncol
2016
;
7
:
108
15
.