More than 11 million Americans are living with a history of cancer (15). Importantly, research is emerging that shows increased weight (e.g., being overweight or obese) has a negative association with disease recurrence, survival, other chronic disease development, and health-related quality of life (HRQoL), particularly in breast cancer survivors (5). On the other hand, research has shown that physical activity (PA) has the opposite effect on the aforementioned outcomes (14,19,20). Taken together, these findings delineate the importance of examining obesity and PA in cancer survivors.
Few studies, however, have examined the association between body mass index (BMI) and PA in cancer survivors. The limited research that is available has shown that active cancer survivors are less likely to be obese or, conversely, obese cancer survivors are less likely to be active (7,16,21). Moreover, although studies have independently examined the associations between BMI, PA, and health-related quality of life (HRQoL) in cancer survivors (7,23), few have examined the potential interaction effects (7,25). Studies in this area have been further limited by small, unrepresentative samples of select cancer survivor groups.
The purpose of the present study was to determine the independent and interactive associations among BMI, PA, and HRQoL in breast, prostate, colorectal, bladder, uterine, and skin melanoma cancer survivors. We hypothesized (purpose 1) that obese cancer survivors would be less likely to be active compared with their nonobese counterparts (7,25). We also hypothesized (purpose 2) that obese cancer survivors would have significantly lower HRQoL compared with nonobese survivors (7,23). Finally, the potential interaction between BMI and PA on HRQoL (purpose 3) was explored. Incidentally, we previously reported the associations between various health behaviors, including PA and HRQoL, from this data set (1). Consequently, the PA-HRQoL relationship is not presented in the current article.
Participants and Procedure
Participants were from the American Cancer Society's (ACS) Study of Cancer Survivors II (SCS-II) (24), a national cross-sectional study of HRQoL among cancer survivors identified through population-based cancer registries. To be included, participants had to be aged 18 yr or older, diagnosed with a local, regional, or distant SEER summary stage cancer (bladder cancer included in situ cancers), a resident of the target state at the time of cancer diagnosis, diagnosed in the calendar year either 2, 5, or 10 yr before sampling, and be diagnosed with female breast, prostate, colorectal, bladder, skin melanoma, or uterine cancer, which were all included in the current article.
A detailed description of the SCS-II methods is reported elsewhere (24). Briefly, overall approval for the study was obtained from the institutional review board of Emory University with additional institutional review board approvals obtained from each participating cancer registry. A total of 36,372 cancer survivors were sampled from 16 cancer registries across the United States. Of these, 2586 survivors were found to be ineligible because of invalid cancer diagnosis or missing cancer stage information, and 4157 were found to be deceased or ineligible because of missing data or out-of-range values on eligibility variables (e.g., age). For the remaining 29,629 survivors, the physician in record was contacted either using an active consent (i.e., were contacted for informed consent to recruit a given survivor) or passive notification (i.e., were sent a letter explaining that recruitment of a given survivor would begin in 3 wk unless study personnel heard from the physician) procedure before approaching the cancer survivor, which resulted in a 91.9% physician consent rate. The remaining 26,802 cancer survivors were invited to participate in the study using an adapted Dillman's tailored design method (9) that included both mail and telephone recruitment data collection procedures. Results showed that 7616 cancer survivors refused to participate (the most common reasons for refusal were lack of interest in research, lack of time, and being too ill to self-report), 2885 were not locatable, and 7196 did not respond, leaving a total of 9105 cancer survivors who agreed to be in the study and completed the questionnaire (i.e., the adjusted overall consent rate was 32.7%) and written informed consent. Only 3306 of the 9105 cancer survivors who responded to the survey were asked questions about their height and weight to calculate BMI because these two variables were added after SCS-II had already commenced. This subsample was analyzed in the current article.
Demographic and clinical information were collected via self-report and registry records. Cancer registry data included date of birth, date of diagnosis, gender, cancer group, and stage of disease at diagnosis. Self-reported data included marital status, education, employment status, income, race, and type of treatment. Survivors were also asked whether they experienced various comorbidities in the past year (i.e., heart problems, hypertension, chronic back pain, arthritis, stroke, osteoporosis, asthma, emphysema or chronic obstructive pulmonary disease, stomach and/or intestinal problems, and diabetes) by having them check all that applied to them. The comorbidities that were checked were summed to obtain an overall comorbidity score.
BMI categories were created on the basis of self-reported weight and height to calculate BMI (kg·m−2). Specifically, cancer survivors were classified into healthy weight (BMI = 18.5-24.9 kg·m−2), overweight (BMI = 25-29.99 kg·m−2), or obese (BMI ≥ 30 kg·m−2) on the basis of the World Health Organization standards (26).
Physical activity was measured via the previously validated Godin Leisure-Time Exercise Questionnaire (LSI) (12,17). The LSI contained three questions, which assessed the frequency of mild, moderate, and strenuous intensity PA performed for at least 10 min in duration during the participants' free time during a typical 1-wk period in the past 3 months. In addition, participants were asked to note the average duration per PA bout that they engaged in within each intensity category. A total LSI score was calculated via the following formula: [(mild frequency × average duration) + (moderate frequency × average duration) + (strenuous frequency × average duration)]. A dummy variable was then created with two categories on the basis of the ACS's recommendation (10) to accumulate at least 150 min of moderate-to-strenuous PA or 60 min of strenuous PA per week (i.e., 0 = did not meet the recommendation; 1 = met recommendation).
Health-related quality of life was measured using the RAND-36 Health Status Inventory. This measure contains four physical domains (i.e., physical functioning, role - physical, bodily pain, and general health), which are weighted and summed to formulate a physical component composite score, and four mental domains (i.e., vitality, social functioning, role - emotional, and mental health), which are weighted and summed to formulate a mental composite score (see the RAND-36 manual for more detail regarding composite calculations ). Both composite scores range from 0 (worst HRQoL score) to 100 (best HRQoL score). The RAND-36 is a well-validated HRQoL measure (13) that has been frequently used in cancer survivors (2,4,11,22).
Data for the 3306 SCS-II survivor subsample was screened, after which descriptive characteristics were generated by cancer type and BMI category. One-way ANOVA and χ2 analyses were conducted to determine whether the demographic/clinical characteristics varied by BMI category within cancer type and whether they were significantly related to PA and/or the physical and mental health composite scores to identify potential confounders for the main analyses (see Tables 1 and 2 for the full list of variables examined). Next, the percentage of cancer survivors within cancer type meeting the PA recommendations was calculated by BMI category, after which unadjusted and adjusted logistic regression analyses were conducted to determine the association between BMI category and meeting the PA guideline (purpose 1). Given that some cell sizes were too small (e.g., n < 3) to conduct a full-factorial ANCOVA (i.e., to detect a potential interaction between BMI category and PA category) within a given cancer type, a custom ANCOVA was conducted to examine the potential main effects of BMI category (healthy weight, overweight, and obese) and PA (met PA recommendation vs did not) simultaneously on HRQoL within each cancer type (purpose 2) followed by Bonferroni post hoc analyses. Given that the PA/HRQoL relationship has already been published in the full SCS-II sample (1), we only present the BMI-HRQoL results in the current article, which were adjusted for PA levels and additional confounders. Finally, to examine the potential interaction between BMI (i.e., treated as a continuous variable to increase power) and PA on HRQoL (purpose 3), a linear regression was performed within the cancer type that entered the demographics/medical covariates identified in the ANCOVA analyses on step 1, the BMI centered around the mean and PA on step 2, and the BMI × PA interaction on step 3.
Data screening procedures resulted in 65 cases being deleted because of advanced cancer stage (n = 38), a BMI < 18.5 kg·m−2 (i.e., underweight: n = 21), or having both (n = 6), leaving a final sample of 3241 cancer survivors (breast = 1013, prostate = 796, colorectal = 668, bladder = 210, uterine = 264, and skin melanoma = 290). The demographic and clinical characteristics by cancer type and BMI category are presented in Tables 1 and 2. As can be seen, several characteristics varied by BMI category across the cancer types. Furthermore, several potential confounding variables were identified within cancer type (see the bottom of Table 3 for PA confounders and the bottom of Table 4 for HRQoL confounders).
Association between BMI category and PA (purpose 1).
As can be seen from Table 3, results showed that compared with healthy-weight cancer survivors, obese breast, prostate, and skin melanoma cancer survivors were significantly less likely to meet the PA guideline. Furthermore, compared with overweight cancer survivors, obese prostate (adjusted odds ratio (OR) = 0.51, 95% confidence interval (CI) = 0.29-0.85) and bladder (adjusted OR = 0.23, 95% CI = 0.08-0.76) cancer survivors were significantly less likely to meet the PA guideline.
Independent associations of BMI category with HRQoL (purpose 2).
For the physical component composite score, there was a significant main effect of BMI category in breast [F(2,709) = 8.00, P = 0.001], prostate [F(2,529) = 5.39, P = 0.005], colorectal [F(2,458) = 6.19, P = 0.002], uterine [F(2,183) = 6.95, P = 0.001], and skin melanoma [F(2,213) = 3.44, P = 0.034] cancer survivors. As can be seen from Table 4, Bonferroni post hoc analyses showed that healthy-weight and/or overweight breast, prostate, colorectal, uterine, and skin melanoma cancer survivors reported significantly better physical functioning compared with their obese counterparts. For the mental health composite score, ANCOVA results showed that there was a significant main effect of BMI category for colorectal cancer survivors [F(2,472) = 3.26, P = 0.040]. Bonferroni post hoc analyses (Table 4) showed that the overweight survivors had significantly better mental health compared with the obese survivors.
BMI × PA interaction effects on HRQoL (purpose 3).
Results from the hierarchical linear regressions showed that none of the BMI × PA interactions was significant for the physical or mental component composite scores across the cancer types.
The first purpose of the present study was to determine whether PA levels varied by BMI (i.e., healthy weight, overweight, and obese). Results showed that BMI had a significant association with PA levels in breast, prostate, bladder, and melanoma cancer survivors. These findings replicated those found in previous breast (16,21) and endometrial (7) cancer survivors and extended them to bladder and skin melanoma cancer survivors. Interestingly, BMI had no association with PA levels in colorectal and uterine cancer survivors. Why the BMI-PA relationship was significant in certain cancers and not in others may be partly due to the different demographic (e.g., 97% of skin melanoma survivors were white compared with 79% of colorectal cancer survivors) and clinical (e.g., colorectal and uterine cancer survivors tended to have a higher number of comorbidities compared with the other cancer types) composition of the cancer types. The importance of these covariates in influencing PA may have also varied across the cancer types. Whatever the case may be, the current finding suggests that BMI may be a determinant of PA levels in breast, prostate, bladder, and skin melanoma cancer survivors. It is unclear form these data, however, if the determinants of PA within BMI categories are different. Research is needed to identify if there are obese-specific correlates of PA in cancer survivors (3) and, if so, to develop interventions accordingly (18).
The second purpose of the present study was to examine the independent association of HRQoL with BMI. Our findings are consistent with previous research in breast (21,23) and endometrial (7) cancer survivors such that BMI was found to be significantly related to the physical composite summary score for breast, prostate, colorectal, uterine, and skin melanoma cancer survivors, but not bladder cancer survivors, independent of PA. The dominant pattern that emerged was that healthy-weight and overweight cancer survivors reported significantly better physical functioning than obese cancer survivors. Given that PA was controlled for in these analyses suggests that factors other than PA are playing an important role in explaining the BMI-physical functioning relationship. Future research needs to identify such factors (e.g., potentially dietary, smoking, fatigue, etc.) to optimize potential intervention effects on physical functioning, particularly in obese cancer survivors.
Results also showed that overweight colorectal cancer survivors had significantly better mental functioning than their obese counterparts. However, the difference would be considered small (d = 0.20) and not clinically meaningful (13). The finding that five of the six cancer types did not have a significant relationship between BMI and mental functioning is consistent with previous research (7). In fact, closer inspection of the means across the different cancer types would suggest that there is little variation across the BMI categories (including colorectal cancer survivors). Of note, however, is that a t score < 38 on the mental health composite score indicates psychological symptoms may be impeding life functioning, whereas a t score > 53 suggests the opposite (13). As can be seen from Table 4, all cancer survivor types are closer to the t score of 53 regardless of BMI category, which is encouraging. As such, it seems that BMI has little association with mental health functioning in cancer survivors. Nonetheless, replication of the current results is necessary using a generic (e.g., the RAND-36) and disease-specific (e.g., the FACT-G) HRQoL measure before any firm conclusions can be drawn.
The third purpose of the present study was to determine whether meeting the ACS's PA recommendation for cancer survivors moderated the BMI-HRQoL relationship. To date, only two studies have empirically addressed this issue in endometrial (7) and ovarian (25) cancer survivors and showed that PA did not moderate the BMI-HRQoL relationship. In line with these findings, this was also the case for all cancer survivor types in the current study. Therefore, a growing literature suggests that BMI and PA contribute independently to the variance in HRQoL among individuals who have been treated for cancer but do not interact to influence HRQoL. This may be interpreted as an encouraging finding because it suggests that an intervention designed to increase PA for healthy weight cancer survivors, for example, may have a similar effect on HRQoL compared with a PA intervention designed for obese cancer survivors. Intuitively, this makes sense because the benefits shown to be influenced by PA (e.g., improved fitness, ameliorated treatment side effects, improved body image, lower depression, etc.) are applicable to all BMI categories (i.e., they are not mutually exclusive). However, a recent randomized controlled trial in lymphoma patients actually found that the effects of aerobic exercise on HRQoL were stronger for healthy weight and obese survivors compared with overweight survivors (8). Consequently, further research is needed on the HRQoL benefits of PA in cancer survivors by weight status.
Despite the importance of this study, there are limitations that need to be considered. First, we used self-report to measure height and weight (i.e., to calculate BMI) and PA. Although this approach is not uncommon in population-based cancer studies (6,7), future studies should attempt to capture these variables by using more objective techniques. Second, the overall response rate for SCS-II (24) was relatively low (i.e., 32.7%), and it suggests that the current study's results may not generalize to all cancer survivors in the population. Therefore, future registry-based cancer survivor studies need to build on the SCS-II methodology to improve the response rate (e.g., by considering face-to-face interviews, incentives, and alternative sampling strategies ). Third, the sample sizes across cancer types, particularly for the bladder and uterine cancer survivors, were not large enough to allow for cancer type comparisons in conjunction with the BMI and PA influences on HRQoL. Therefore, future studies should overcome this limitation to allow for statistical comparisons to be made. Finally, the current article collapsed colon and rectal cancer survivors into one group. Future studies should determine whether potential differences exist among the BMI, PA, and HRQoL relationships between these particular cancer types.
The present study provides several novel findings that are of interest. In particular, the percentage of cancer survivors meeting the ACS's PA guideline seems to vary by BMI in breast, prostate, bladder, and skin melanoma cancer survivors. In addition, BMI has an independent association with HRQoL; however, the interactive association of BMI and PA on HRQoL was negligible. Importantly, clarifying the relationship between BMI and PA across different cancer types will help identify potential target groups for future PA interventions that will help ameliorate the negative side effects of cancer and improve HRQoL.
This study was funded by the American Cancer Society. The ACS SCS were funded as an intramural program of research conducted by the ACS Behavioral Research Center. The authors acknowledge the cooperation and efforts of the cancer registries and public health departments from the states of Arizona, California (regions 2-6), Colorado, Delaware, Illinois, Iowa, Maine, Massachusetts, Michigan, Nebraska, New Jersey, Pennsylvania, Washington, and Wyoming. The authors also thank the staff of the hundreds of hospitals that reported cases to the participating cancer registries. Lastly, the authors thank the thousands of cancer survivors, their physicians, and their loved ones who contributed to the collection of these data. The authors assume full responsibility for analyses and interpretation of these data. Chris Blanchard and Kerry Courneya are supported by the Canada Research Chairs Program. Results from the present study do not constitute endorsement by American College of Sports Medicine.
1. Blanchard C, Courneya K, Stein K. Cancer survivors' adherence to lifestyle behavior recommendations and associations with health-related quality of life: results from the American Cancer Society's Studies of Cancer Survivors (SCS-II
). J Clin Oncol
2. Blanchard C, Stein K, Baker F, et al. Association between current lifestyle behaviors and health-related quality of life in breast, colorectal, and prostate cancer survivors. Psychol Health
3. Blanchard CM, McGannon KR, Spence JC, et al. Social ecological correlates of physical activity in normal weight, overweight, and obese individuals. Int J Obes (Lond)
4. Camilleri-Brennan J, Steele R. Prospective analysis of quality of life and survival following mesorectal excision for rectal cancer. Br J Surg
5. Chlebowski R, Aiello E, McTiernan A. Weight loss in breast cancer patient management. J Clin Oncol
6. Coups E, Ostroff J. A population-based estimate of the prevalence of behavioural risk factors among adult cancer survivors and noncancer controls. Prev Med
7. Courneya K, Karvinen K, Campbell K, et al. Associations among exercise, body weight, and quality of life in a population-based sample of endometrial cancer survivors. Gynecol Oncol
8. Courneya K, Sellar C, Stevinson C, et al. Moderator effects in a randomized controlled trial of exercise training in lymphoma patients. Cancer Epidemiol Biomarkers Prev
9. Dillman D. Alternative questionnaire delivery: in person, to groups, and through publications. In: Mail and Internet Surveys: The Tailored Design Method
. New York (NY): John Wiley & Sons Inc; 2000. p. 245-62.
10. Doyle C, Kushi L, Byers T, et al. Nutrition and physical activity during and after cancer treatment: an American Cancer Society guide to informed choices. CA Cancer J Clin
11. Ganz P, Rowland J, Desmond K, Meyerowitz B, Wyatt G. Life after breast cancer: understanding women's health-related quality of life and sexual functioning. J Clin Oncol
12. Godin G, Shepard RJ. A simple method to assess exercise behavior in the community. Can J Appl Sport Sci
13. Hays R. Appendix B: computation of scale and composite scores for the RAND-36 HSI. In: RAND-36 Health Status Inventory
. San Antonio (TX): The Psychological Corporation, Harcourt Brace and Company; 1998. p. 61-2.
14. Holmes M, Chen W, Feskanich D, Kroenke C, Colditz G. Physical activity and survival after breast cancer diagnosis. JAMA
15. Horner M, Ries L, Krapcho M, et al. SEER Cancer Statistics Review, 1975-2006
[Internet]. Bethesda (MD): National Cancer Institute; [cited 2009 March 5]. Available from: http://seer.cancer.gov/csr/1975_2006/
16. Irwin M, McTiernan A, Bernstein L, et al. Physical activity levels among breast cancer survivors. Med Sci Sports Exerc
17. Jacobs DR, Ainsworth BE, Hartman TJ, Leon AS. A simultaneous evaluation of ten commonly used physical activity questionnaires. Med Sci Sports Exerc
18. Karvinen K, Courneya K, Campbell K, et al. Correlates of exercise motivation and behavior in a population-based sample of endometrial cancer survivors: an application of the theory of planned behavior. Int J Behav Nutr Phys Act
19. Meyerhardt J, Giovannucci E, Holmes M, et al. Physical activity and survival after colorectal cancer diagnosis. J Clin Oncol
20. Meyerhardt J, Heseltine D, Niedzwiecki D, et al. Impact of physical activity on cancer recurrence and survival in patients with stage III colon cancer: findings from CALGB 89803. J Clin Oncol
21. Milne H, Gordon S, Guilfoyle A, Wallman K, Courneya K. Association between physical activity and quality of life among western Australian breast cancer survivors. Psychooncology
22. Penson D, Stoddard M, Pasta D, Lubeck D, Flancders S, Litwin M. The association between socioeconomic status, health insurance coverage, and quality of life in men with prostate cancer. J Clin Epidemiol
23. Peuckmann V, Ekholm O, Rasmussen N, et al. Health-related quality of life in long-term breast cancer survivors: nationwide survey in Denmark. Breast Cancer Res Treat
24. Smith T, Stein K, Mehta C, et al. The rationale, design, and implementation of the American Cancer Society's Studies of Cancer Survivors. Cancer
25. Stevinson C, Faught W, Steed H, et al. Associations between physical activity and quality of life in ovarian cancer survivors. Gynecol Oncol
26. World Health Organization. Diet, Nutrition and the Prevention of Chronic Diseases: Technical Report Series 916
[Internet]. Geneva (Switzerland): World Health Organization; [cited 2009 March 5]. Available from: http://whqlibdoc.who.int/trs/who_TRS_916.pdf
Keywords:©2010The American College of Sports Medicine
BODY MASS INDEX; HRQoL; PA; SCS-II