Dietary supplement use among elderly, long-term cancer survivors

Abstract

Introduction

The purpose of the present study was to assess dietary supplement use and its association with micronutrient intakes and diet quality among older (≥65y), long-term survivors (≥5 years post-diagnosis) of female breast, prostate, and colorectal cancer.

Methods

The sample included 753 survivors who participated in telephone screening interviews to determine eligibility for a randomized diet and physical activity intervention trial entitled RENEW: Reach-out to ENhancE Wellness in Older Cancer Survivors. Telephone surveys included two 24-hour dietary recalls and items regarding supplement use (type, dose, and duration). Nutrient intakes were compared to Dietary Reference Intakes (DRIs). Diet quality was assessed using the revised Healthy Eating Index (HEI). Descriptive statistics and multivariate logistic regression were used in this cross-sectional study.

Results

A majority of survivors (74%) reported taking supplements, with multivitamins (60%), calcium/vitamin D (37%), and antioxidants (30%) as the most prevalent. Overall proportions of the total sample with dietary intakes below Estimated Average Requirements (EARs) were substantial, although supplement users had more favorable mean HEI scores (P<0.01) and nutrient intakes for 12 of the 13 vitamins and minerals investigated (P-values <0.05). Supplement use was positively associated with older age (≥70 years) (odds ratio (OR) 1.70; 95% confidence interval (95% CI) 1.17, 2.46) and female gender (OR 1.49; 95% CI 1.04, 2.13), and negatively associated with current smoking (CI 0.40, 95% CI 0.21, 0.76). Individuals scoring higher on the Total Fruit (OR 1.12, CI 1.01, 1.23), Whole Grain (OR 1.14, CI 1.04, 1.25), and Oil (OR 1.10, CI 1.01, 1.11) components of the HEI were significantly more likely to take supplements, while those scoring higher on the Meat and Beans category (OR 0.81, CI 0.71, 0.93) were significantly less likely to take supplements. Compared to those with less than a high school education, survivors with a professional or graduate degree were significantly more likely to use supplements (OR 2.18, CI 1.13, 4.23).

Discussions/Conclusions

Demographic, disease, and health-related correlates of supplement use follow similar trends observed in the general population as well as previous reports from other cancer survivor populations. Supplement use may reduce the prevalence of nutrient inadequacies in this population, though survivors who use supplements are the least likely to need them.

Implications for Cancer Survivors

Supplement use may be an effective means for many survivors to achieve adequate nutrient intakes; however, open communication between healthcare providers and survivors is needed to ensure potential concerns are addressed as supplement use may not always be beneficial.

Introduction

Cancer survival rates are rising along with the population of older adults in the United States [1], leading to an expanding population of elderly (≥65y), long-term cancer survivors, a group at increased risk for progressive, recurrent, and secondary cancer, osteoporosis, obesity, cardiovascular disease, diabetes, and functional decline [2-5]. Elderly cancer survivors may initiate diet, physical activity, and other lifestyle changes post-diagnosis in an attempt to prevent recurrence or other chronic disease, or improve overall health and quality of life [2, 3, 5, 6]. One of the most common changes in behavior among cancer survivors is the use of new dietary supplements [5], a broad category of products including vitamins, minerals, herbs or other botanicals, amino acids, glandular extracts, or other non-nutrient ingredients [7]. Despite widespread use of dietary supplements, their role in promoting health and preventing cancer recurrence, secondary cancers, and other chronic diseases remains unclear. According to the National Institutes of Health (NIH) [8], the state of the scientific evidence is insufficient to establish firm recommendations concerning the use of multivitamin-multimineral (MVMM) supplements for chronic disease prevention in the general population.

Nevertheless, several potential beneficial roles for dietary supplements have been suggested. A daily MVMM supplement supplying nutrients at or below recommended intake levels is generally considered an effective approach to increase the prevalence of nutrient intake adequacy among adults [9, 10]. Older adults in particular may have compromised nutrient intakes resulting from a variety of factors [11], including but not limited to functional or financial constraints [12, 13], acute and chronic illnesses, medication use [14], oral health concerns [15], or various environmental barriers [13]. Diminished absorption and utilization of nutrients concomitant with aging or medication use also may exacerbate existing intake inadequacies [16]. In addition to potential benefits of MVMM use, calcium and vitamin D supplementation have been found to slow bone loss associated with cancer treatment [17] as well as aging [18].

At the other end of the spectrum, a growing body of evidence suggests that the use of various dietary supplements in some cases could pose health-related risks [19-28]. For example, a recent investigation found an increased risk of advanced and fatal prostate cancer among older men who used MVMM supplements more than seven times per week compared to never users [28]. Overall, very little is known concerning the efficacy and safety of use of most dietary supplements [29]. The use of supplements may result in nutrient intakes surpassing the Tolerable Upper Intake Levels (ULs) established by the Institute of Medicine of the National Academy of Sciences (IOM/NAS) [30-33]. However, few studies have explored the use of dietary supplements among elderly, long-term cancer survivors.

Characterizing supplement use patterns is an important step in accurately evaluating the associated health benefits and risks of supplement use. With older age as a leading risk factor for cancer and compromised nutrient intakes more likely among older adults, a better understanding of dietary supplement use among elderly, long-term cancer survivors is of particular interest. Further, as financial resources may be limited among older adults [12, 13], dietary supplement and food purchasing decisions become increasingly important. The primary aim of this study was to assess dietary supplement use and its association with micronutrient intakes and diet quality within a population of older (≥65y), long-term survivors (≥5 years post-diagnosis) of female breast, prostate, and colorectal cancer.

Methods

Study sample

Participants included 753 older survivors (≥65y) at least 5 years removed from a diagnosis of female breast, prostate, or colorectal cancer who completed extensive screening measures for a randomized controlled trial entitled Reach out to ENhancE Wellness in Older Cancer Survivors (RENEW), a study designed to test the efficacy of a home-based diet and physical activity intervention aimed at improving the physical functioning of older survivors. The Institutional Review Boards at both Duke University Health System and the Pennsylvania State University approved the research protocol. Survivors were identified for this study via the North Carolina Central Cancer Registry (NCCCR) and previously established physician referral networks, described in detail elsewhere [34]. To be considered eligible for this trial, individuals also had to be community-dwelling, English-speaking and writing, overweight (BMI ≥25 kg/m²) but not morbidly obese (BMI <40 kg/m²), with no pre-existing medical or physical conditions that would preclude adherence to an unsupervised diet and physical activity program (i.e., unstable angina, congestive heart failure, end-stage renal disease, severe orthopedic conditions, etc).

Eligible individuals participated in two telephone interviews administered by the Diet Assessment Center at the Pennsylvania State University. All individuals who completed the telephone interviews as a part of the screening process for RENEW were included in the exploratory analysis (n=753) discussed here, which included survivors of female breast (n = 321), prostate (n = 319), and colorectal (n = 113) cancer residing across North America. Since individuals who reported exercising 150 minutes or more per week were excluded from participation, only 641 of these survivors ultimately were enrolled in RENEW. Chi-square analyses revealed that men were more likely to screen out of RENEW as a result of reporting 150 minutes or more of moderate to vigorous physical activity each week. No other statistically significant differences were noted between the RENEW trial sample and the full sample discussed here.

Demographic, disease, and health-related characteristics

Demographic and disease-related data including the type of cancer, date of diagnosis, cancer stage, age, race, and gender were provided by the NCCCR for a large majority of cases and by the oncology care physician for self-referred individuals. More detailed participant information was collected by interviewers during the two baseline computer-assisted telephone interviews. Questions pertaining to the presence of co-morbid conditions, the type of cancer treatment, body weight, height, physical functioning, and general health or quality of life perceptions were asked in conjunction with two 24-hour dietary recalls. Physical activity data were ascertained via the telephone-adapted Community Health Activities Model Program (CHAMPS) questionnaire, a previously validated tool for use among older adults [35].

Dietary and supplemental intake

Dietary intake and supplement use data were obtained from two unannounced 24-hour recalls performed by trained interviewers at the Diet Assessment Center at the Pennsylvania State University. Previous research has shown 24-hour dietary recalls performed over the telephone to be a practical and valid data collection tool [36]. Interviewers collected dietary intake data using the interactive Nutrition Data System-Revised (NDS-R) software (NCC Food and Nutrient Database System Version 2006, Nutrition Coordinating Center, Minneapolis, MN). During the dietary recall interviews, participants were asked to indicate if they had taken any dietary supplements the previous day. Individuals reporting the use of at least one supplement on one of the two 24-hour dietary recalls were defined as supplement users, while those who did not indicate the use of any dietary supplement were defined as nonusers. Supplement users were asked to list any supplements taken the previous day and the duration of use for each supplement. Product name was collected for MVMM supplements and other combination preparations, and specific dose was recorded for the remaining supplements reported, including single vitamin, mineral, and herbal supplements.

Dietary supplements were grouped into 12 categories: 1) MVMM products, 2) single calcium, calcium-containing antacids, calcium combination supplements, and single vitamin D, 3) antioxidants (carotenoids, vitamin C, vitamin E, selenium, and related combinations), 4) herbal and other botanical supplements, 5) other single vitamins, 6) other single minerals, 7) other vitamin combinations, 8) other mineral combinations, 9) fatty acids, fish oils, and other oils, 10) glucosamine, chondroitin, and related combinations, 11) amino acids, nucleic acids, and proteins including co-enzymes and 12) fiber supplements (various types).

Several conversions to standardized units were made before adding nutrient values from supplements to dietary intakes. Milligrams of supplemental folic acid were multiplied by a factor of 1.7 to convert to dietary folate equivalents (DFE). Supplemental vitamin A and beta-carotene recorded in International Units (IU) were multiplied by factors of 0.3 and 0.15, respectively, to convert to retinol activity equivalents (RAE). Supplemental vitamin E recorded in IU was considered to be the synthetic form, all rac-α-tocopherol, and was converted to α-tocopherol equivalents (TE) by a factor of 0.45. Vitamin D recorded in IU was divided by a factor of 40 to determine the equivalent amount in micrograms (mcg). Intake from MVMM formulations, eye-health antioxidant combination formulas, and general antioxidant combination supplements were estimated by using default nutrient profiles of the most commonly reported products: Centrum Silver (Wyeth Consumer Healthcare, Madison, NJ), Ocuvite PreserVision AREDS Formula for Macular Degeneration (Bausch & Lomb, Rochester, NY), and Spring Valley Antioxidant Formula with Minerals (Nature’s Bounty, Inc, Bohemia, NY).

Overall diet quality was measured by the revised Healthy Eating Index (HEI) [37], a food-derived assessment tool consisting of twelve individual scored components that sum to a maximum score of 100 (best) and a minimum score of zero (worst). The following six components have a maximum score of five: Total Fruit, Whole Fruit, Total Vegetables, Dark Green and Orange Vegetables and Legumes, Total Grains, and Whole Grains. Milk, Oil, Meat and Beans, Sodium, and Percent Calories from Saturated Fat have a maximum score of 10. The remaining component of Calories from Solid Fat, Alcohol, and Added Sugar has a maximum score of 20.

Comparisons to Dietary Reference Intakes

Nutrient intake estimates were generated in three ways: 1) from dietary intake alone for nonusers of supplements; 2) from dietary intake alone for users of supplements; and 3) from dietary combined with supplemental intake for users. The Estimated Average Requirement (EAR) cut-point method was used to estimate the prevalence of nutrient intake inadequacy [38]. With the cut-point method, the prevalence of inadequate nutrient intake is the proportion of the population with usual intake levels below the average requirement (EAR). Estimates of proportions of individuals meeting Adequate Intakes (AIs) were generated for nutrients without EARs established by the IOM/NAS; however, these estimates were not used to assess the prevalence of nutrient inadequacy, as AIs were not developed for this purpose [38]. Rather, the proportion of our sample meeting the AI for each respective vitamin or mineral without an EAR was considered to have nutrient intakes that were most likely adequate. Comparisons to ULs were used to estimate the proportion of survivors at risk for adverse health effects from elevated nutrient intakes [30-33].

Statistical analyses

Dietary supplement users were defined as participants reporting the use of at least one supplement on one of the two 24-hour dietary recalls. Student’s t-tests and chi-square tests were used to compare supplement users and nonusers on demographic, lifestyle, and health-related characteristics. Multivariate logistic regression was performed to identify significant predictors of supplement use. The dependent variable of supplement use was dichotomous (use/nonuse). All characteristics significantly associated with supplement use from bivariate chi-square analyses were entered as candidate predictor variables in regression analysis (i.e., age, gender, cancer type, education, tobacco use, total HEI score, and individual HEI component scores of interest). A P-value of ≤ 0.05 was considered statistically significant for all statistical tests. All data were analyzed using SAS version 9.1.3 (2007, SAS Institute Inc, Cary, NC).

Results

Sample description

A total of 753 individuals were included in the final sample. Study participants were predominantly white, well-educated, and from moderate income households. Age ranged from 65 to 87 years, with a median of 73 years. The mean years from diagnosis of survivors at the time of the telephone screening was 8.5 ± 2.6 years. Slightly less than one-half of participants reported no moderate or vigorous physical activity each week. Additional demographic, disease and health-related characteristics of the total sample are presented in Table 1.

Table 1.

Relation of demographic, disease, and health-related characteristics with dietary supplement use by older cancer survivors (n = 753)

Characteristics	Total no. of participants	% participants reporting supplement use
	n	%
Age
65-69y	231	69*
70+	522	77*
Gender
Female	394	78**
Male	359	70**
Ethnicity
White	669	75
Nonwhite	84	69
Cancer Type
Female breast	321	80*
Colorectal	113	69*
Prostate	319	71*
Education
< High school	56	64*
High school/GED	217	73*
Some college/college degree	356	74*
Professional/graduate degree	121	81*
Income
≤ $30,000	219	71
>$30,000	467	76
No response	67	73
Stage at diagnosis
Localized	523	75
Regional	203	74
Unknown	27	67
BMIa
Overweight (25-30 kg/m2)b	464	75
Obese (≥30 kg/m2)	289	72
Tobacco use
Current smoker	46	52**
Non-smoker	707	76**

^*P < 0.05

^**P < 0.01

^aBMI = body mass index; calculated as kg/m²

^bIncludes 11 individuals with BMIs from 22-24 because BMI was assessed twice as a part of a two tiered screening process; second measure was used

Supplement use

Dietary supplement use was common in our study (74%). Use of supplements was more prevalent among breast cancer survivors, as well as non-smoking survivors and individuals with higher levels of educational attainment (P-values <0.05). Table 2 presents the frequency of use by categories of supplements. MVMM products were the most commonly reported supplement, followed by calcium/vitamin D, antioxidants, fatty acids/oils, glucosamine/chondroitin, and herbal/botanical supplements. The median number of formulations reported was two, with a range from one to 25 formulations. Ninety percent of supplement users reported using at least one formulation for greater than one year, 54% for five or more years, and 19% for ten or more years.

Table 2.

Specific categories of supplements reported by supplement users, according to cancer site

Supplement categories	All supplement users (n = 559)		Breast cancer (n = 256)		Colorectal cancer (n = 78)		Prostate cancer (n = 225)
	n	%	n	%	n	%	n	%
MVMM	446	80	200	78	62	79	184	82
Calcium, vitamin D, and combinations	277	50	167	65	40	51	90	22
Antioxidantsa	231	41	98	38	32	41	101	45
Herbal/botanical	99	18	42	16	13	17	44	20
Other single vitamins	91	16	47	18	9	12	35	16
Other single minerals	52	9	28	11	5	6	19	8
Other single vitamin combinations	41	7	15	6	5	6	21	9
Other single mineral combinations	1	0.2	1	0.4	0	0	0	0
Fatty acids/fish oils/other oils	160	29	66	26	21	27	73	31
Amino acids/nucleic acids/proteins	30	5	13	5	1	1	16	7
Glucosamine, chondroitin, and combinations	106	19	38	15	16	20	52	23
Fiber (various types)	35	6	12	5	7	9	16	7

^aIncludes vitamins C and E, carotenoids, selenium, and related combinations

Comparisons to Dietary Reference Intakes

Comparisons of the proportions of nonusers versus users below EARs prior to the addition of nutrients from dietary supplements are shown in Table 3 (supplement nonusers vs. supplement users: dietary sources only), along with comparisons of the proportions of nonusers versus users after consideration of nutrients from supplements (supplement nonusers vs. supplement users: dietary + supplemental intake). Although smaller proportions of supplement users compared to nonusers had dietary intakes below EARs, overall proportions of the total sample with dietary intakes below EARs for the following vitamins and minerals were substantial: vitamin E (82%), magnesium (80%), vitamin C (51%), vitamin A (48%), vitamin B6 (43%), zinc (42%), and folate (28%). Consideration of micronutrients from supplements substantially reduced proportions of users below EARs, most notably for vitamins A, B6, C, and E, magnesium, and zinc.

Table 3.

Nutrient intakes and proportion of survivors below Estimated Average Requirements (EARs) and above Tolerable Upper Intake Levels (ULs)

	EAR male/female	Mean intake ± SD	Below EAR		Above UL
			n	%	n	%
Vitamin A (RAE)	625/500
Supplement nonusersa		690 ± 626	110	57df	1	0.5
Supplement users: dietary intake onlyb		703 ± 639	251	45d	5	1
Supplement users: dietary + supplemental intakec		1576 ± 937	47	8f	5	1
Vitamin E (alpha-tocopherol equivalents in mg)	12/12
Supplement nonusersa		7.6 ± 5.8	167	86f	0	0
Supplement users: dietary intake onlyb		10.1 ± 11.4	451	81	0	0
Supplement users: dietary + supplemental intakec		86 ± 125	67	12f	0	0
Vitamin C (mg)	75/60
Supplement nonusersa		63 ± 48	124	64ef	0	0
Supplement users: dietary intake onlyb		80.5 ± 53.8	261	47e	0	0
Supplement users: dietary + supplemental intakec		292 ± 379	52	9f	4	0.5
Vitamin B6 (mg)	1.4/1.3
Supplement nonusersa		1.5 ± 0.7	91	47f	0	0
Supplement users: dietary intake onlyb		1.7 ± 0.9	229	41	0	0
Supplement users: dietary + supplemental intakec		8.0 ± 19.1	42	8f	13	2
Vitamin B12 (mcg)	2.0/2.0
Supplement nonusersa		5.4 ± 7.9	33	17f	No UL
Supplement users: dietary intake onlyb		4.8 ± 4.7	91	16
Supplement users: dietary + supplemental intakec		99 ± 369	18	3f
Folate (dietary folate equivalents (DFE))	320/320
Supplement nonusersa		443 ± 235	67	35df	0	0
Supplement users: dietary intake onlyb		498 ± 285	147	26d	1	0.2
Supplement users: dietary + supplemental intakec		1078 ± 505	22	4f	22	4
Magnesium (mg)	350/265
Supplement nonusersa		222 ± 85	174	90ef	0	0
Supplement users: dietary intake onlyb		246 ± 95	428	77e	0	0
Supplement users: dietary + supplemental intakec		338 ± 138	225	40f	10	2
Zinc (mg)	9.4/6.8
Supplement nonusersa		11.9 ± 17.4	92	48f	7	4
Supplement users: dietary intake onlyb		10.0 ± 6.4	222	40	4	1
Supplement users: dietary + supplemental intakec		25.5 ± 16.9	37	7f	48	9
Iron (mg)	6.0/5.0
Supplement nonusersa		12.8 ± 5.9	9	5	0	0
Supplement users: dietary intake onlyb		13.8 ± 6.9	6	1	4	0.5
Supplement users: dietary + supplemental intakec		15.1 ± 17.3	6	1	8	1.5

^aSupplement nonusers (n = 194); only dietary sources were used to calculate nutrient intake

^bSupplement users: dietary intake only (n = 559); supplemental intake was excluded from nutrient intake

^cSupplement users: dietary + supplemental intake (n = 559); supplemental intake was included in nutrient intake

^dSignificant differences noted between users and nonusers (dietary intake only); P < 0.05

^eSignificant differences noted between users and nonusers (dietary intake only); P < 0.001

^fSignificant differences noted between users and nonusers (dietary + supplemental intake); P < 0.001

Table 4 presents proportions of survivors meeting AIs for vitamins D and K, calcium, and potassium, four micronutrients for which available data were insufficient or inconclusive for the IOM/NAS to determine EARs. Comparisons of nonusers and users prior to the consideration of dietary supplements are presented, along with comparisons after the addition of nutrients to dietary intake estimates. Very small proportions of both users and nonusers met AIs from diet alone. Consideration of micronutrient contributions from supplements considerably increased proportions of users meeting AIs for calcium and vitamin D, but not vitamin K or potassium.

Table 4.

Nutrient intakes and proportion of survivors meeting Adequate Intakes (AIs) and above Tolerable Upper Intake Levels (ULs)

	AI male/female	Mean intake ± SD	Below EAR		Above UL
			n	%	n	%
Vitamin D (mcg)	15/15
Supplement nonusersa		4.3 ± 5.2	9	5f	1	0.5
Supplement users: dietary intake onlyb		4.3 ± 4.1	16	3	1	0.2
Supplement users: dietary + supplemental intakec		15.1 ± 7.8	306	55f	3	0.5
Vitamin K (mcg)	120/90
Supplement nonusersa		102 ± 131	50	26de	No UL
Supplement users: dietary intake onlyb		113 ± 115	198	35d
Supplement users: dietary + supplemental intakec		121 ± 115	216	39e
Calcium (mg)	1,200/1,200
Supplement nonusersa		594 ± 289	186	4f	0	0
Supplement users: dietary intake onlyb		668 ± 350	33	6	2	0.5
Supplement users: dietary + supplemental intakec		1156 ± 593	205	37f	17	3
Potassium (mg)	4,700/4,700
Supplement nonusersa		2161 ± 793	0	0	No UL
Supplement users: dietary intake onlyb		2288 ± 781	6	1
Supplement users: dietary + supplemental intakec		2368 ± 799	9	1.5

^aSupplement nonusers (n = 194); only dietary sources were used to calculate nutrient intake

^bSupplement users: dietary intake only (n = 559); supplemental intake was excluded from nutrient intake

^cSupplement users: dietary + supplemental intake (n = 559); supplemental intake was included in nutrient intake

^dSignificant differences noted between users and nonusers (dietary intake only); P < 0.05

^eSignificant differences noted between users and nonusers (dietary + supplemental intake); P < 0.01

^fSignificant differences noted between users and nonusers (dietary + supplemental intake); P < 0.001

Overall, small proportions of users had intakes exceeding ULs (Tables 3 and 4). Approximately 10% of users exceeded the UL for zinc, while 4% or less of users exceeded the ULs for vitamins A, B6, C, and D, folic acid, calcium, iron, and magnesium. No individuals exceeded the UL for vitamin E, and to date, no ULs have been established by the IOM/NAS for vitamins B12 and K, or potassium.

Factors associated with supplement use

All characteristics significantly associated with supplement use in bivariate analyses (e.g. age, gender, cancer type, education, tobacco use, total HEI score, and individual HEI component scores of interest) were entered as candidate predictor variables in multivariate logistic regression analysis. As shown in Table 5, results indicated that supplement use was positively associated with older age (≥70 years) (OR 1.70; 95% CI 1.17, 2.46) and female gender (OR 1.49; 95% CI 1.04, 2.13), and negatively associated with current smoking (CI 0.40, 95% CI 0.21, 0.76). Individuals scoring higher on the Total Fruit (OR 1.12, CI 1.01, 1.23), the Whole Grain (OR 1.14, CI 1.04, 1.25), and the Oil (OR 1.10, CI 1.01, 1.11) components of the HEI were significantly more likely to take supplements, while those scoring higher on the Meat and Beans category (OR 0.81, CI 0.71, 0.93) were significantly less likely to take supplements. Compared to those with less than a high school education, survivors with a professional or graduate degree were significantly more likely to use supplements (OR 2.18, CI 1.13, 4.23).

Table 5.

Factors associated with supplement use in multivariate logistic regression

	Odds Ratioa	95% Confidence Intervals
Age
65-69y	1.00
≥ 70y	1.70	(1.17, 2.46)
Gender
Males	1.00
Females	1.49	(1.04, 2.13)
Education
< High school	1.00
High school/GED	1.39	(0.79, 2.45)
Some college/college degree	1.47	(0.90, 2.51)
Professional/graduate degree	2.18	(1.13, 4.23)
Tobacco Use
Non-smoker	1.00
Current smoker	0.40	(0.21, 0.76)
HEIb Components
Total Fruit	1.12	(1.01, 1.23)
Whole Grain	1.14	(1.04, 1.25)
Meat and Beans	0.81	(0.71, 0.93)
Oil	1.10	(1.01, 1.11)

^aOdds ratios control for all other variables presented in table

^bHEI = Healthy Eating Index

Discussion

In this sample of older, long-term cancer survivors, a majority (74%) reported taking dietary supplements, ranging from 70% among survivors of colorectal cancer to 80% among breast cancer survivors. The prevalence found in this study is comparable to other previously reported estimates of supplement use among cancer survivors [39], but notably higher than estimates of supplement use among healthy older adults aged 60 years and above (63%) (1999-2000 NHANES) [40]. While the observation of greater prevalence of use among older adults and cancer survivors has been reported previously, our study is the first to target older survivors at least five years removed from a cancer diagnosis.

A wide variety of supplements were taken by participants, although MVMM products (60%), calcium/vitamin D (37%), antioxidants (30%), fatty acids/oils (21%), and glucosamine/chondroitin (14%) were more prevalent. The three most commonly reported supplements in our study are similar to the most frequently reported supplements in prior investigations [40-44]; however, larger proportions of survivors in our study used fatty acids/oils and glucosamine/chondroitin.

Similar to findings from other surveys and studies [41, 45-48], our data suggest that supplement use is associated with several demographic, disease, and health-related characteristics, and it may serve as a marker for a range of other health-related behaviors, evidenced by lower prevalence rates of smoking and higher scores on measures of diet quality. While our study is the first to explore associations between the revised HEI [37] and supplement use, other investigations have found supplement users to have higher diet quality scores [49, 50] using the prior HEI [51]. Associations found between the food-derived HEI and supplement use support previous findings revealing higher intakes of fruits, vegetables, and fiber, and lower intakes of saturated fat by supplement users [6, 47, 49]. Although supplement use has tracked with higher levels of physical activity and lower rates of overweight and obesity in previous investigations [40, 41, 52], supplement users and nonusers in our study presented with similar distributions of BMI and physical activity. In light of the initial RENEW screening measures ensuring that older adults in our study were either overweight or obese, as well as the previously observed association of higher BMI levels and lower levels of physical activity among older adults [53, 54], some of the more active cancer survivors were likely excluded from our sample.

Consideration of micronutrient contributions from dietary supplements substantially reduced proportions of individuals below EARs for most micronutrients, including vitamins A, C, E, and B6, folate, magnesium, and zinc, paralleling findings from previous investigations [9, 10, 46]. Dietary intakes of vitamin B12 and iron were sufficient for a vast majority of survivors before consideration of dietary supplements. It is important to note that despite improvements in magnesium intake after consideration of supplemental intake, 40% of users remained below the EAR.

Overall, very few individuals met AIs for calcium, vitamin D, potassium, and vitamin K from diet alone. Supplemental intake considerably increased proportions of supplement users above AIs for calcium (6% to 37%) and vitamin D (3% to 55%), but did not sizably increase proportions above AIs for potassium (1% to 2%) or vitamin K (35% to 39%). These findings are not surprising as single supplementation with potassium or vitamin K was rare and combination formulas used by our participants contained low levels of both micronutrients. Low values of vitamin K in MVMM products targeted for older adults may reflect the known drug-nutrient interactions of vitamin K with anticoagulant medication [32], while low levels of potassium may be attributable to the potential risk of serious adverse reactions associated with high intakes by individuals with impaired renal function or by those on potassium-sparing diuretics [55]. Overall, our data indicate that MVMM supplements could be better formulated to reduce the risk of inadequate intakes, with special consideration of magnesium, calcium, and vitamin D. Increasing levels of vitamin K or potassium in MVMM supplements should be cautioned as subpopulations of older adults are at increased risk of drug-nutrient interactions and adverse health effects from high intakes.

Despite the increasing availability of fortified foods and the high prevalence of supplement use, a majority of supplement users in our sample did not reach intake levels deemed excessive by the IOM/NAS. Only 4% of supplement users in our study exceeded the UL for folic acid from supplements and fortified foods; therefore, risk of high folic acid intakes masking vitamin B12 deficiencies [56] may not be a significant concern among the majority of older cancer survivors. Similarly, only 3% of supplement users exceeded the UL for calcium (>2500 mg). Nevertheless, guidelines recently issued by the World Cancer Research Fund and the American Institute for Cancer Research cautioned that calcium intakes of 1500 mg or more per day may be associated with elevated risk of aggressive prostate cancer [57]. Our findings revealed 27% of prostate cancer survivors who used dietary supplements had calcium intakes above 1500 mg per day, thus suggesting the need for increased guidance within this subgroup of cancer survivors. Indeed, the finding that only 5% of our sample met the AI for calcium from diet alone suggests that while supplementation might still be necessary for many older adults, special consideration regarding appropriate dose for prostate cancer survivors is warranted.

Approximately 10% of supplement users reported zinc intakes above the UL, which is comparable to an estimate from a previous investigation among older adults [9]. Prolonged consumption of high intakes of zinc has been associated with decreased high-density lipoprotein cholesterol, reduced copper status, and a weakened immune response [32]. Further, a recent study observed an increased risk of fatal prostate cancer among older men who used a zinc supplement in addition to an MVMM supplement more than seven times per week [28]. However, nearly one-half of the individuals who did not use supplements failed to meet the EAR for zinc; thus, a substantial proportion of elderly cancer survivors may benefit from supplemental zinc. Our data indicate that while the use of an MVMM product with a modest amount of zinc may benefit a sizable proportion of older adults, the use of an individual zinc supplement or an antioxidant combination product supplying high doses of zinc may pose potential health risks.

There are several strengths of the present study. The size of our sample and the distribution of survivors among three cancer types are two notable strengths compared to previous investigations of supplement use patterns among survivors, which tend to be limited to breast cancer survivors [46, 58-61] or are comprised of smaller samples [5, 46, 58, 59, 62]. Despite consistently high supplement use rates observed across studies, supplemental intakes rarely are included in nutrient intake estimations, which likely results in underestimating total intake [63]. Further, the two telephone surveys included open-ended questions pertaining to supplement use, which allowed interviewers to capture components of combination supplements that might have been missed with close-ended questions (yes/no).

As in all studies, there are several limitations with the current investigation. One limitation to consider relates to sample and respondent bias. Our sample included a self-selected population of long-term, elderly survivors, who were predominantly white, well-educated, and interested in participating in a home-based diet and exercise program. Survivors who did not respond to the mailed study invitation may have differed from respondents with regards to demographic and lifestyle characteristics, as well as overall health, which may have influenced our findings. Given the variations in MVMM product composition [63], the use of a default nutrient profile likely overestimated supplemental intake for some individuals, while underestimating intake for others. The degree to which this biased our results is uncertain, as the error may have been randomly distributed. In addition, the true prevalence rate of dietary supplement use may be greater than estimated in our study because episodically consumed supplements could have been missed with only two dietary recalls. Lastly, the examination of multiple associations may have resulted in findings reaching statistical significance by chance alone.

Results from the present study indicate a high prevalence of dietary supplement use among elderly, long-term survivors of female breast, prostate, and colorectal cancer who are interested in lifestyle modification and health promotion. Elderly cancer survivors represent a rapidly growing population; thus, a better understanding of dietary supplement use and health-related behavior practices can provide valuable insight for the development of behavioral interventions and recommendations aimed at improving health outcomes among survivors [2]. Lastly, it is important to consider individual variability when developing recommendations concerning supplement use, as no single supplement is optimal for all older cancer survivors. Both individualized assessment and sound scientific evidence of efficacy and safety of product use should provide the foundation for the development of recommendations regarding dietary supplement use in this population.