Physical activity was directly associated with physical functioning (p<.001) in this sample. When accounting for potential mediating variables such as recent falls and comorbidities, the associations were only modestly attenuated. Neither sex nor age modified the association with similarly strong associations seen in men and women, and in those both under and over the age of 80 years. Moderate-vigorous intensity activity appeared to be the important component of total activity, as light activity alone was not associated with physical function. Activity of many types, including housework, gardening, walking, and exercise appeared to be related to better function; however, stretching was not.
Participants reported engaging in 2773 ± 2586 kcals/wk of physical activity, on average, however, the median level of physical activity reported was 2135 kcal/wk. Stewart, et al. [37
] previously reported a mean energy expenditure of 2420 (SD, 1831) as measured by the CHAMPS questionnaire in a sample of men and women aged 65–90 yrs, which appears reasonably consistent with our findings. This volume of activity appears quite high given that most adults do not meet the 1000 kcal/wk recommendation by the CDC/ACSM [45
]. To some extent, this potential overreporting could be explained by the inclusion of light intensity activities on the CHAMPS questionnaire; however, even the moderate-vigorous activity reported appears quite high. Regardless, if the potential over-reporting is consistent across activity levels it should not affect our ability to compare low vs. high activity groups. Another important possibility though, is that we have selection bias in our sample and that we were more likely to receive responses from survivors who do more physical activity. Conclusions regarding the general population of older colorectal cancer survivors must thus be tempered with this potential selection bias in mind.
The mean Physical Functioning (PF) subscale score of our participants (55.0 ± 30.5) is close to the SF-36 age-based population norms of men and women aged ≥75 years (53.20 ± 29.98). A previous study of long-term female colorectal cancer survivors in Wisconsin reported a mean PF score of 62.4 ± 25.8 for a group of 81 women aged 65–85 years [47
], which again appears consistent with the results of this study when considering the differences in age and sex. We observed a very strong and significant 22-point difference in function between people reporting the lowest vs. the highest levels of total physical activity, and a 21-point difference across levels of moderate-vigorous activity. This difference is more marked than the 15.4 point difference between regular vigorous exercisers and non-regular vigorous exercisers previously reported by Demark-Wahnefried, et al.[23
] in a group of breast and prostate cancer survivors at ~18 months post-diagnosis. This inconsistency may be due to the difference in physical activity assessment, with inclusion of moderate activity in our moderate-vigorous score vs. the focus on vigorous activity in the study by Demark-Wahnefried et al.[23
]. Consequently, their non-regular vigorous exercisers may have included people doing more moderate activity, thus reducing the potential difference in function. Importantly, our study suggests the largest difference in physical function between two consecutive activity groups is seen between the 1st and 2nd quartiles. This suggests that movement from the least active group to just a modest activity level is associated with the largest change in physical function. Additionally, we saw very strong relationships with housework, gardening, and walking, suggesting that exercise, per se, may not be necessary to maintain physical function at higher levels.
] has argued that a 5–10 point difference on any of the eight SF-36 subscales is clinically relevant, thus, the 10 and 12-point differences observed between the least active group and the published age norm and the overall sample mean, respectively, suggest a clinically meaningful functional impairment in these individuals. Conversely, the most active men and women report PF levels nearly 12 and 10 points higher than the published age norm and sample mean, respectively. In essence, these most active individuals report functional status similar to that of individuals aged 10 years younger [42
], further reinforcing the notion that physical activity and functional ability are highly related in older adults.
To our knowledge, this is the first study to examine sex differences in the association of physical activity and function among cancer survivors. We hypothesized that women would show a stronger association between physical activity and function, considering that women tend to report both more numerous and more severe limitations than do men [4
]. Though women did report lower mean functional status at all levels of activity, the association of physical activity and function did not vary by sex. We also hypothesized that the physical activity and function association may be modified by age, and that some of our oldest subjects would not have the same, direct relationship. We divided our subjects into two groups based on the median age of our sample of 80.7 years, and found the same strong relationship among those aged 65–80.7 as among those aged 80.7–103 years of age. Our data suggest that physical activity may be beneficial at any age.
One benefit of observational studies over intervention studies is the ability to examine the natural distribution in types and amounts of activity that are done in given populations. In particular, it is useful to know if non-exercise activities such as housework or gardening are associated with better physical function. We used the CHAMPS physical activity questionnaire to categorize responses to specific questions into general categories of activities that people frequently report participating in such as housework, gardening, and walking. When mutually adjusted, all three of these categories of activity, in addition to more traditionally examined exercise activities, were significantly associated with better function. Importantly, for housework, gardening, and exercise, people doing any activity of those types had significantly better function when compared to those who did none. With walking, function improved with increasing amounts of activity. Perhaps not too surpisingly, stretching was unrelated to physical function.
], resistance-based [28
], aerobic-based [25
], and combined aerobic and resistance-based [29
] exercise interventions have been conducted in breast, prostate and colorectal cancer survivors. While the majority of these interventions have shown significant improvements in functional measures [25
], no study has compared the effects of resistance vs. aerobic activities on physical function in cancer survivors. We hypothesized that resistance activities would be more strongly associated with function than aerobic activities, assuming that loss of strength following treatment, particularly related to surgery and/or chemotherapy, would be more functionally limiting than loss of aerobic capacity. However, no differences in function were seen based on the type of activity after controlling for other energy expenditure and relevant covariates. This may be a result of examining a composite physical function score rather than looking at task-specific function (e.g. carrying a load vs. walking ¼ mile). Cancer treatment did not appear related to physical function in our analyses suggesting that there may not be long-term consequences of either chemotherapy or radiation treatment in colorectal cancer patients who have lived this long. Given the potential errors in our treatment measures, which were not directly taken from medical records, further investigation into this issue is warranted.
Strengths of this study include a large sample size, including both men and women who had the same primary cancer. We used validated measures of both physical activity and function. Additionally, the use of the CHAMPS physical activity questionnaire allowed us to examine the potential difference between intensity and types of activity. Limitations of this study include the cross-sectional design, self-reported measures of activity and function, and limited generalizability. Given the study design, we cannot attribute higher levels of function to physical activity participation; however, activity and function remained strongly associated even after including several potential markers of lower functional status (eg. comorbidities, falls, hospitalizations, pain) in the model, suggesting some independent relation of activity and function. As with all self-report surveys, our measures may be subject to recall bias, and as previously mentioned, we may also have some selection bias. Finally, the modest response rate (65%) accompanied by the lack of racial and ethnic diversity in this sample limits the generalizability of our findings.
Though residual confounding cannot be ruled out, and the results of this study do not overtly suggest that physical activity improves function in older cancer survivors, taken in context with data from physical activity intervention trials [24
] and other population-based studies [22
], the overall evidence suggests that physical activity may be one avenue for improving physical function in long-term cancer survivors. Our results suggest that the potential benefits would be apparent for both men and women, for people choosing to participate in a variety of activities, and more so for participation in moderate-vigorous activities than light intensity activities. Further studies are needed to identify characteristics of a physical activity intervention program that would be safe, promote high levels of adherence, and maximize functional benefits among older cancer survivors.