- Email: [email protected]
Dietary supplement use patterns in men with prostate cancer: the Cancer Prostate Sweden Study
original article
Annals of Oncology 22: 967–972, 2011 doi:10.1093/annonc/mdq456 Published online 6 October 2010
Dietary supplement use patterns in men with prostate cancer: the Cancer Prostate Sweden Study A. Westerlund1, G. Steineck1,2, K. Ba¨lter3, P. Stattin4, H. Gro¨nberg3 & M. Hedelin1,2* 1
Division of Clinical Cancer Epidemiology, Department of Oncology, Institute of Clinical Sciences, Sahlgrenska Academy at University of Gothenburg, Gothenburg; Division of Clinical Cancer Epidemiology, Department of Oncology and Pathology, Karolinska Institutet, Stockholm; 3Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm; 4Department of Surgery and Perioperative Sciences, Urology and Andrology, Umea˚ University Hospital, Umea˚, Sweden 2
Received 1 March 2010; revised 13 April 2010; accepted 22 July 2010
controls in Sweden. Age-adjusted binary regression with an identity link was carried out to estimate prevalence differences and corresponding 95% confidence intervals (CIs). Modifying effects of lifestyle- and diet-related factors were explored by statistical assessment of additive interaction. Results: Among men with PCa, 542 individuals (48%) had used supplements, which was a 10% (95% CI: 5.9%– 15%) higher prevalence than among population controls. Among individuals with high intake of fatty fish, vegetables, and phytoestrogens, but low intake of saturated fat, supplement use was 29% (95% CI: 18%–41%) more common in men with PCa than in population controls. We found no evidence of heterogeneity by categories of education, smoking history, body mass index, fiber, fruit, or phytoestrogen intake, treatment, or disease stage. Conclusion: Supplement use is common in Swedish men with PCa, especially among those with a healthy dietary pattern. Key words: complementary therapies, dietary supplements, prostatic neoplasms
introduction The use of complimentary and alternative medicine (CAM) is common among cancer patients, with reported prevalences ranging between 36% and 70% [1–3]. Of the various types of CAM that cancer patients use, dietary supplements represent a considerable fraction [3, 4]. In general, total CAM use (or use of dietary supplements specifically) in cancer patients or longer term survivors is associated with higher educational level and higher income and in some instances with younger or older age, smoking habits, or disease stage [1–5]. Previous studies indicate that supplement use is associated with dietary habits [6, 7]. Most studies of CAM use among men with prostate cancer (PCa) have been conducted in North American men [5, 7–9]. To our knowledge, European studies are few, relatively small and none has compared CAM use among men with PCa and matched population controls [10–12]. We explored whether men with PCa were more likely to use dietary supplements than men without PCa in a population-based study of Swedish men. In addition, we investigated whether supplement use
*Correspondence to: Dr. M. Hedelin, Division of Clinical Cancer Epidemiology, Department of Oncology, Institute of Clinical Sciences, Sahlgrenska University Hospital, SE-413 45 Gothenburg, Sweden. Tel: +46-70-7401858; Fax: +46-31-820114; E-mail: [email protected]
was dependent on factors such as dietary habits, body mass index (BMI), smoking, education, cancer treatment, or disease status.
materials and methods study population Cancer Prostate in Sweden is a population-based case–control study, with enrollment from January 2001 to September 2002 [13]. Men, 35–79 years of age, with pathologically verified adenocarcinoma of the prostate (International Classification of Diseases-10: C61) were reported to four regional cancer registries in Sweden. Population controls were randomly selected from the Swedish Population Registry and frequency matched by age and geographic residence. Men with advanced PCa disease were defined as those with at least one of the following criteria: tumor–node–metastasis stage, = T3/T4, N+, M+; Gleason score = 8–10; or prostate-specific antigen (PSA) level ‡ 100 ng/ml [14]. Men with localized PCa disease were those not meeting any of the above criteria. Of the 1895 men with PCa who were invited to the study, 1499 (79%) agreed to participate by completing the questionnaire. Of 1684 invited control subjects, 1130 (67%) completed the questionnaire [13]. The investigation was
ª The Author 2010. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For permissions, please email: [email protected]
original article
cancer (PCa) and to what extent lifestyle, disease or other factors influence such use.
Patients and methods: We evaluated supplement use in 1127 men with incident PCa and in 900 population
Downloaded from http://annonc.oxfordjournals.org/ at Florida Atlantic University on July 29, 2015
Background: In a European setting, we know little about the use of dietary supplements among men with prostate
original article approved by the Ethics Committees at Karolinska Institutet and at Umea˚ University.
statistical methods Men with PCa who stated that they had used supplements the current year or that they had initiated supplement use the same year as they were diagnosed were defined as current users. The median time between diagnosis and return of the questionnaire to the study center was 5.6 months (5%–95% percentiles: 2–12 months). Population controls who stated that they had used dietary supplements the current year or that they had initiated supplement use the same year as the questionnaire was returned to the study center, also were defined as current users. Individuals who stated in the questionnaire that they had used supplements the current year but gave information on use before that during the telephone interviews were defined as former users. In the analyses, only current and never users were included, while former users (n = 602) were omitted. In total, 1127 men with PCa and 900 population controls were included in the analyses. Baseline characteristics of cases and controls were compared using a two-sided t-test for continuous variables and a v2 test for categorical variables. In exploring whether individuals with PCa were more likely to use dietary supplements (current versus never user) than subjects without PCa, we used ageadjusted (with age in 5-year intervals) binary regression with an identity or logarithmic link summarized in terms of prevalence differences and prevalence ratios, respectively, and corresponding 95% confidence intervals (CIs). PCa (yes/no) was considered the exposure and dietary supplement use (current user/never user) the outcome (event).
968 | Westerlund et al.
Supplements were divided into subgroups: antioxidants (e.g. vitamins C and E, selenium, beta-carotene, Oxigard, coenzyme Q10), vitamins and/or minerals (e.g. multivitamins with or without minerals, calcium, folate, magnesium, B vitamins, zinc), omega-3 fatty acids, supplements for symptoms of benign prostatic hyperplasia (e.g. pollen preparations, saw palmetto, and field pumpkin preparations), and other supplements (e.g. algae, garlic preparations, ginseng, ginkgo biloba, and other natural remedies). The subcategories serve only as descriptive data as we could not determine whether individuals who had used more than one type of supplement had done so before or after diagnosis (the current year or before that for controls) or both. The association between dietary supplement use and PCa was stratified by lifestyle- and diet-related categories, including education, smoking, BMI, intake of fatty fish, saturated fat, vegetables, fiber, fruit, and food items rich in phytoestrogens. We also considered possible differences in supplement use between men with PCa who had received treatment (n = 813) and those who had not (n = 314) and between those with localized (n = 628) or advanced disease (n = 448) (unknown status, n = 51). Intake of the various foods was categorized into tertiles or quartiles based on the distribution among the population controls. Based on previous results regarding dietary habits generally considered as healthy [13, 17, 18], we created a variable for increasing ‘health index’, which grouped the men according to intake of fatty fish, vegetables, food items rich in phytoestrogens, and saturated fat. The index had three levels: (i) low intake of fatty fish, food items rich in phytoestrogens, and vegetables and high intake of saturated fat; (ii) intermediary intake; and (iii) high intake of fatty fish and food items rich in phytoestrogens and vegetables and low intake of saturated fat. To explore modifying effects of lifestyle- and diet-related factors, statistical assessment of interaction effects was carried out on the additive effect scales. PCa status was represented by two indicator variables comparing men with PCa against men without PCa, and each lifestyle- and dietrelated factor was categorized into three or four groups (see Table 2). Interaction was assessed by the same product terms under a linear odds model, and all analyses were adjusted for age. Analyses were carried out using the STATA System Software, version 10.1.
results characteristics of study participants Baseline characteristics, intake of macronutrients, and use of supplements among the study participants are presented in Table 1. We found no statistically significant differences between subjects with or without PCa, respectively, with regard to BMI, smoking history, level of education, intake of main groups of macronutrients, or overall energy intake. The mean age among men without PCa was somewhat higher than among men with PCa. With one exception, men with PCa were more likely to use supplements than population controls in each dietary supplement subcategory. There was no statistically significant
Volume 22 | No. 4 | April 2011
Downloaded from http://annonc.oxfordjournals.org/ at Florida Atlantic University on July 29, 2015
assessment of food and dietary supplement intake The questionnaire included questions on education, smoking history, treatment type, weight, height, and a validated foodfrequency questionnaire to measure average intake of foods, beverages, and dietary supplements the past 12 months [15]. The respondents were asked if they had used any vitamins, minerals, or other dietary supplements during this time period. They answered by ‘No, I have never eaten any dietary supplements’; ‘Yes, I have eaten dietary supplements before, but not the current year’; or ‘Yes, I have eaten dietary supplements the current year’. Men who stated that they had used supplements the current year or before that were asked to further specify which type(s) of supplements they had used (e.g. multivitamins with or without minerals, vitamin C, vitamin E, selenium, calcium, or fish oil/omega-3 fatty acids). Respondents also could state ever use of specific supplements, such as Cernitol (pollen extract), Curbicin (extracts from saw palmetto fruit and field pumpkin seeds), Echinacea (purple coneflower), garlic, ginseng, coenzyme Q10, and different trademarks of vitamins and minerals. If information was missing from the mailed questionnaire, participants were contacted by phone to complete the missing information. Questionnaire data about the intake of food items were converted into intake of energy and nutrients by linkage to the nutrient database created by the Swedish National Food Administration [16].
Annals of Oncology
original article
Annals of Oncology
Table 1. Selected characteristics of participants with questionnaire data in the Cancer Prostate Sweden Study Characteristic
Men with PCa, n = 1127
Men without PCa, n = 900
0.01 0.46
0.07
0.54
– 0.1
0.1 0.2 0.1
<0.01
0.01
<0.01
<0.01
0.2
0.01
BMI, body mass index; PCa, prostate cancer.
difference in use of omega-3 fatty acid supplements between men with and without PCa (data not shown).
PCa status and use of dietary supplements Five hundred and forty-two men (48%) with PCa and 341 men (38%) without PCa had used any supplement(s) the previous year (Table 2). The use of supplements was 9.9% (95% CI: 5.6%–14.2%) more common in men with PCa than in population controls. Table 2 also shows the prevalence
Volume 22 | No. 4 | April 2011
discussion In this study, we found that dietary supplement use was more common among men with PCa than in population controls. In the group that had the healthiest dietary intake pattern (high intake of fatty fish, vegetables, and food items rich in phytoestrogens and low intake of saturated fat), a higher proportion of men with PCa were supplement users compared with population controls. A handful of studies of CAM use among men with PCa report the prevalence of dietary supplement use specifically. Accordingly, 15%–56% of men with PCa use dietary supplements [3, 5, 8, 9, 19]. In a study that focused exclusively on supplement use among men with PCa, the user prevalence was 73% [7]. Possible explanations for the differences in prevalence rates compared with our data include differing definitions of dietary supplements and differing points in time
doi:10.1093/annonc/mdq456 | 969
Downloaded from http://annonc.oxfordjournals.org/ at Florida Atlantic University on July 29, 2015
Age, mean 6 SD (years) 67 6 7.3 68 6 7.5 BMI, mean 6 SD (kg/m2) 26 6 3.4 26 6 3.4 Education, n (%) Compulsory school, 0–9 532 (47) 423 (47) years Upper secondary, 10–12 431 (38) 372 (41) years University, ‡13 years 161 (14) 99 (11) Missing 3 (1) 6 (1) Smoking history, n (%) Never 444 (39) 341 (38) Ever 667 (59) 542 (60) Missing 16 (2) 17 (2) Treatment, n (%) Hormones 493 (44) – Surgery 248 (22) – Radiotherapy 72 (6) – No treatment 214 (28) 900 Total energy intake, 9245 8913 median (kJ) Proportion of energy intake, median (%) from: Fat (%) 32 32 Protein (%) 16 16 Carbohydrate (%) 50 49 Use of any dietary supplement, n (%) Never 585 (52) 559 (62) Current 542 (48) 341 (38) Use of antioxidants, n (%) Never 840 (75) 715 (79) Ever 287 (25) 185 (21) Use of vitamins and/or minerals, n (%) Never 731 (65) 667 (74) Ever 396 (35) 233 (26) Use of supplements for benign prostatic hyperplasia, n (%) Never 998 (89) 849 (94) Ever 129 (11) 51 (6) Use of omega-3 fatty acids, n (%) Never 998 (89) 814 (90) Ever 129 (11) 86 (10) Use of other supplements, n (%) Never 850 (75) 724 (80) Ever 277 (25) 176 (20)
P value
differences and prevalence ratios of dietary supplement use comparing PCa status, stratified by dietary intakes, BMI, education, and smoking history. When carrying out interaction analysis, we found a significant interaction between PCa status and saturated fat intake (P = 0.01), vegetable intake (P = 0.04), and healthy dietary intake pattern (health index) (P = 0.002) and borderline significance for fatty fish intake (P = 0.06) but not for level of education, smoking history, BMI, or fiber, fruit, or phytoestrogen intake (Table 2). In the group with the healthiest dietary intake pattern, a higher proportion of men with PCa was supplement users compared with population controls (prevalence difference: 29%, 95% CI: 18%–41%). Regardless of PCa status, we found that the proportion of supplement users was higher among men with high educational level (‡10 years) compared with men with low educational level (0–9 years) (prevalence difference: 19, 95% CI: 15%–24%) (data not shown). But we found no evidence of heterogeneity by categories of education (P = 0.86) (Table 2). The use of supplements was more common among younger men than older men, irrespective of PCa status. The difference in prevalence of supplement use comparing men aged 40–59 years with men aged 69–74 years was 13% (95% CI: 5.3%–21%) (data not shown). Furthermore, men who were overweight or obese (BMI > 25) used supplements to a 7% lower degree than normal-weight (BMI < 25) men (95% CI: 212% to 23%) (data not shown). There was no significant difference in rate of overall supplement use between men who had received some kind of treatment (hormones, radiotherapy, or surgery) and those who had not received any treatment (prevalence difference: 3.6, 95% CI: 22.9% to 10%) (data not shown). When specific supplement use was stratified by type of treatment, men who had received hormone therapy were more likely to use omega-3 supplements compared with those who had not received any treatment (prevalence difference: 6.2, 95% CI: 1.8%–11%) (data not shown). Supplement user rates did not significantly differ between men with advanced compared with localized disease (prevalence difference: 22.5, 95% CI: 28.7% to 3.6%) (data not shown). Adjustment for potential confounders (education level, smoking history, etc.) did not substantially change any of the results (data not shown).
original article
Annals of Oncology
Table 2. Prevalence differences and prevalence ratios with corresponding 95% CI for dietary supplement use by levels of PCa status, stratified by education, smoking, BMI, and dietary intake Characteristic
PCa statusa
Dietary supplementb use, n (%) Never Current
All participants
No Yes
559 (62) 585 (52)
341 (38) 542 (48)
0 (ref.) 9.9 (5.6–14.2)
1 (ref.) 1.3 (1.1–1.4)
306 329 207 205 41 49
(72) (62) (56) (48) (41) (30)
117 203 165 226 58 112
(28) (38) (44) (52) (59) (70)
0 9.5 0 8.2 0 11
(ref.) (3.6–15) (ref.) (1.3–15) (ref.) (21.2 to 23)
1 1.4 1 1.2 1 1.2
(ref.) (1.1–1.6) (ref.) (1.0–1.4) (ref.) (0.97–1.4)
0.86
211 221 338 356
(62) (50) (38) (53)
130 223 130 223
(38) (50) (38) (50)
0 12 0 9.1
(ref.) (4.9–19) (ref.) (3.5–15)
1 1.3 1 1.2
(ref.) (1.1–1.5) (ref.) (1.1–1.4)
0.47
186 205 272 278 76 77
(58) (47) (63) (54) (64) (58)
133 235 159 236 42 56
(42) (53) (37) (46) (36) (42)
0 11 0 8.7 0 6.7
(ref.) (4.1–18) (ref.) (2.4–15) (ref.) (25.3 to 19)
1 1.3 1 1.2 1 1.2
(ref.) (1.1–1.5) (ref.) (1.1–1.4) (ref.) (0.88–1.6)
0.22
32 55 467 489 60 41
(60) (60) (62) (53) (61) (39)
21 37 281 440 39 65
(40) (40) (38) (47) (39) (61)
0 0.43 0 9.5 0 21
(ref.) (216 to 17) (ref.) (4.8–14) (ref.) (7.1–35)
1 1.1 1 1.3 1 1.5
(ref.) (0.72–1.6) (ref.) (1.1–1.4) (ref.) (1.1–2.0)
0.06
159 118 134 150 132 173 134 144
(70) (47) (60) (55) (59) (50) (60) (55)
68 131 91 125 92 170 90 116
(30) (53) (40) (45) (41) (50) (40) (45)
0 22 0 4.8 0 7.8 0 5.1
(ref.) (14–31) (ref.) (23.9 to 14) (ref.) (20.40 to 16) (ref.) (23.7 to 14)
1 1.7 1 1.1 1 1.2 1 1.1
(ref.) (1.4–2.2) (ref.) (0.91–1.4) (ref.) (1.0–1.4) (ref.) (0.90–1.4)
0.01
131 131 136 141 138 146 154 161
(60) (55) (60) (53) (61) (52) (68) (49)
88 107 90 125 90 137 73 167
(40) (45) (40) (47) (39) (48) (32) (51)
0 4.5 0 7.4 0 9.0 0 19
(ref.) (24.4 to 14) (ref.) (21.3 to 16) (ref.) (0.42–18) (ref.) (10–27)
1 1.1 1 1.2 1 1.2 1 1.6
(ref.) (0.89–1.3) (ref.) (0.99–1.5) (ref.) (1.0–1.5) (ref.) (1.3–2.0)
0.04
166 183 150 138 126 137 117 127
(74) (61) (67) (61) (56) (50) (52) (39)
59 115 75 90 99 139 108 198
(26) (39) (33) (39) (44) (50) (48) (61)
0 11 0 4.5 0 7.1 0 13
(ref.) (3.3–19) (ref.) (24.2 to 13) (ref.) (21.6 to 16) (ref.) (4.7–22)
1 1.4 1 1.1 1 1.2 1 1.3
(ref.) (1.1–1.9) (ref.) (0.84–1.4) (ref.) (0.98–1.4) (ref.) (1.1–1.5)
0.93
970 | Westerlund et al.
Prevalence ratioc (95% CI)
P value for homogeneity
Volume 22 | No. 4 | April 2011
Downloaded from http://annonc.oxfordjournals.org/ at Florida Atlantic University on July 29, 2015
Education level (years) 0–9 No Yes 10–12 No Yes >13 No Yes Smoking history Never No Yes Ever No Yes BMI <25 No Yes 25 to >30 No Yes ‡30 No Yes Fatty fish intake Never No Yes £2 times/week No Yes >3 times/week No Yes Saturated fat intake (g/d) 0–26 No Yes 26–33 No Yes 33–44 No Yes 44–154 No Yes Vegetable intake (g/d) 0–184 No Yes 185–245 No Yes 245–327 No Yes 327–1768 No Yes Fiber intake (g/d) 0–18 No Yes 18–23 No Yes 23–29 No Yes 29–76 No Yes
Prevalence differencec (95% CI)
original article
Annals of Oncology
Table 2. (Continued) Characteristic
PCa statusa
Fruit intake (g/d) 0–74
Prevalence differencec (95% CI)
Prevalence ratioc (95% CI)
P value for homogeneity
139 126 137 147 125 146 158 160
(61) (51) (61) (53) (57) (53) (69) (51)
88 122 87 128 94 131 72 155
(39) (49) (39) (47) (43) (47) (31) (49)
0 10.4 0 7.3 0 5.8 0 17
(ref.) (1.5–19) (ref.) (21.4 to 16) (ref.) (23.0 to 15) (ref.) (8.8–25)
1 1.3 1 1.2 1 1.1 1 1.6
(ref.) (1.0–1.6) (ref.) (0.96–1.5) (ref.) (0.91–1.3) (ref.) (1.3–2.0)
0.36
142 174 130 132 152 140 135 133
(59) (54) (61) (54) (67) (53) (61) (47)
98 148 83 114 75 126 85 148
(41) (46) (39) (46) (33) (47) (39) (53)
0 4.9 0 7.3 0 14 0 14
(ref.) (23.4 to 13) (ref.) (21.8 to 16) (ref.) (5.2–22) (ref.) (4.8–22)
1 1.1 1 1.2 1 1.4 1 1.4
(ref.) (0.93–1.4) (ref.) (0.96–1.5) (ref.) (1.1–1.8) (ref.) (1.1–1.7)
0.58
84 109 401 424 74 52
(64) (60) (62) (54) (62) (33)
48 72 248 362 45 108
(36) (40) (38) (46) (38) (68)
0 2.9 0 7.8 0 29
(ref.) (28.0 to 14) (ref.) (2.7–13) (ref.) (18–41)
1 1.1 1 1.2 1 1.8
(ref.) (0.83–1.5) (ref.) (1.1–1.4) (ref.) (1.4–2.3)
0.002
a
Men with pathologically verified adenocarcinoma of the prostate (International Classification of Diseases-10:C61) are defined as ‘yes’ and population controls as ‘no’. b Dietary supplement use refers to use after PCa diagnosis (for men with PCa) or the current year (for population controls). Dietary supplements include, e.g. multivitamins, single vitamins and minerals, omega-3 fatty acids, and natural remedies. c Adjusted for age. d Health index: (i) low intake of fatty fish, food items rich in phytoestrogens and vegetables, high intake of saturated fat; (ii) intermediary intake; and (iii) high intake of fatty fish and food items rich in phytoestrogens and vegetables and low intake of saturated fat. BMI, body mass index; CI, confidence interval; PCa, prostate cancer; SD, standard deviation.
in relation to PCa diagnosis when use was measured, as well as varying places of residence of the participants. Regardless of PCa status, we found that men with higher educational level were more likely to use supplements than men without high school education and that younger men were more likely to use supplements than older men. These findings are in line with several studies of CAM use among survivors of PCa as well as other cancers [3, 5–7, 9, 10, 19] and in the general population [20]. Furthermore, our observation that a healthy dietary intake pattern influences the use of supplements finds support in previous investigations. For example, PCa and other cancer survivors with high intake of fruits and vegetables or of fruit, whole grain, and oil had a higher use of supplements [6, 7]. Comparable results were found in a healthy population [20]. Overall dietary supplement use did not differ between men with PCa who had received treatment and those who had not. This is consistent with a smaller study in which no difference in CAM use was found between treatment groups [21]. Wilkinson et al. [8] and Boon et al. [19] did, though, find that men who chose hormonal therapy were more likely to use CAM
Volume 22 | No. 4 | April 2011
compared with those who chose surgery or radiation and CAM nonusers, respectively. We found that men who had received hormonal therapy were more prone to use omega-3 fatty acid supplements compared with men who had not received any treatment. Wiygul et al. [7] found that men who chose other treatment options than surgery or radiation were more prone to use herbals, while in the study by Diefenbach et al. [9], watchful waiting was the only treatment option associated with use of herbals. In terms of disease stage, we found no difference in supplement use between men with localized and advanced disease. Advanced [19] and progressive [8] disease have previously been positively associated with CAM use. Strengths of our study include its population-based design, large size, and complete and rapid case ascertainment. The ethnic homogeneity of our study population reduces the risk of confounding by population stratification. The Swedish cancer registries record almost 100% of all incident cases and since there was limited PSA testing in our study population at the time of enrollment [13], our results pertain mainly to nonPSA-detected clinically significant PCa. Any problem with
doi:10.1093/annonc/mdq456 | 971
Downloaded from http://annonc.oxfordjournals.org/ at Florida Atlantic University on July 29, 2015
No Yes 74–135 No Yes 135–220 No Yes 220–1430 No Yes Phytoestrogen intake (g/d) 0–8 No Yes 9–23 No Yes 23–41 No Yes 41–596 No Yes Health indexd 1 No Yes 2 No Yes 3 No Yes
Dietary supplementb use, n (%) Never Current
original article
funding Swedish Cancer Society (050158 to HG, 080062 to MH); Region Va¨stra Go¨taland, Sweden.
acknowledgements The authors would like to thank all study participants in the CAPS study; Ulrika Lund Unde´n for skillfully coordinating the study center at Karolinska Institutet; Lotta Spa˚ngberg, Berit Andersson, and Britt Eriksson for conducting thorough interviews; and all urologists whose patients were included in the CAPS study.
972 | Westerlund et al.
disclosure The authors declare no conflicts of interest.
references 1. Molassiotis A, Fernadez-Ortega P, Pud D et al. Use of complementary and alternative medicine in cancer patients: a European survey. Ann Oncol 2005; 16: 655–663. 2. Paltiel O, Avitzour M, Peretz T et al. Determinants of the use of complementary therapies by patients with cancer. J Clin Oncol 2001; 19: 2439–2448. 3. Patterson RE, Neuhouser ML, Hedderson MM et al. Types of alternative medicine used by patients with breast, colon, or prostate cancer: predictors, motives, and costs. J Altern Complement Med 2002; 8: 477–485. 4. Velicer CM, Ulrich CM, Velicer CM, Ulrich CM. Vitamin and mineral supplement use among US adults after cancer diagnosis: a systematic review. J Clin Oncol 2008; 26: 665–673. 5. Chan JM, Elkin EP, Silva SJ et al. Total and specific complementary and alternative medicine use in a large cohort of men with prostate cancer. Urology 2005; 66: 1223–1228. 6. Miller P, Demark-Wahnefried W, Snyder DC et al. Dietary supplement use among elderly, long-term cancer survivors. J Cancer Surviv 2008; 2: 138–148. 7. Wiygul JB, Evans BR, Peterson BL et al. Supplement use among men with prostate cancer. Urology 2005; 66: 161–166. 8. Wilkinson S, Gomella LG, Smith JA et al. Attitudes and use of complementary medicine in men with prostate cancer. J Urol 2002; 168: 2505–2509. 9. Diefenbach MA, Hamrick N, Uzzo R et al. Clinical, demographic and psychosocial correlates of complementary and alternative medicine use by men diagnosed with localized prostate cancer. J Urol 2003; 170: 166–169. 10. Wilkinson S, Farrelly S, Low J et al. The use of complementary therapy by men with prostate cancer in the UK. Eur J Cancer Care 2008; 17: 492–499. 11. Ponholzer A, Struhal G, Madersbacher S. Frequent use of complementary medicine by prostate cancer patients. Eur Urol 2003; 43: 604–608. 12. Salmenpera¨ L. The use of complementary therapies among breast and prostate cancer patients in Finland. Eur J Cancer Care 2002; 11: 44–50. 13. Hedelin M, Klint A, Chang ET et al. Dietary phytoestrogen, serum enterolactone and risk of prostate cancer: the cancer prostate Sweden study (Sweden). Cancer Causes Control 2006; 17: 169–180. 14. Sobin LH, Wittekind C (eds): TNM Classification of Malignant Tumours (UICC), 6th edition. New York: Wiley-Liss 2002. 15. Chang ET, Smedby KE, Zhang SM et al. Dietary factors and risk of non-hodgkin lymphoma in men and women. Cancer Epidemiol Biomarkers Prev 2005; 14: 512–520. 16. The National Food Administration’s Food Database. http://www7.slv.se/ Naringssok/SokLivsmedel.aspx (12 August 2010, date last accessed). 17. Hedelin M, Chang ET, Wiklund F et al. Association of frequent consumption of fatty fish with prostate cancer risk is modified by COX-2 polymorphism. Int J Cancer 2007; 120: 398–405. 18. Jakobsen MU, O’Reilly EJ, Heitmann BL et al. Major types of dietary fat and risk of coronary heart disease: a pooled analysis of 11 cohort studies. Am J Clin Nutr 2009; 89: 1425–1432. 19. Boon H, Westlake K, Stewart M et al. Use of complementary/alternative medicine by men diagnosed with prostate cancer: prevalence and characteristics. Urology 2003; 62: 849–853. 20. Foote JA, Murphy SP, Wilkens LR et al. Factors associated with dietary supplement use among healthy adults of five ethnicities: the multiethnic cohort study. Am J Epidemiol 2003; 157: 888–897. 21. Lippert MC, McClain R, Boyd JC, Theodorescu D. Alternative medicine use in patients with localized prostate carcinoma treated with curative intent. Cancer 1999; 86: 2642–2648.
Volume 22 | No. 4 | April 2011
Downloaded from http://annonc.oxfordjournals.org/ at Florida Atlantic University on July 29, 2015
misclassification of disease status (in this study considered as the exposure of interest) therefore is unlikely. Prevalences of dietary supplement use may be influenced by erroneous reports by the respondents, measurement errors are unavoidable. Measurement errors may arise, for example, if the cancer patients were more prone to state their supplement use than the controls. They may, for example, have a greater interest in the possible effect of dietary supplementation on the disease. If the cancer patients overreported their use, this could account for the apparent positive association between PCa status and supplement use. On the other hand, some underreporting is also possible among patients who think they should use more supplements than they currently do. However, most participants (1787 of 2027) were contacted by phone to complete missing information from the questionnaire, including information about supplement use. The interviewers were blinded to the respondents’ PCa status. Therefore, we do not believe that this type of potentially differential misclassification has had any significant impact on our results. Finally, the study was not originally designed to measure whether use of dietary supplements had been initiated or changed after diagnosis. Even though we have information on when supplement use started, our definition is not optimal and we probably have some misclassification. Prior supplement use could lead to misclassification of our binary outcome variable, but such misclassification is likely non-differential and would attenuate any true association. In conclusion, use of dietary supplements among Swedish men with PCa is widespread and even higher in men with a healthy dietary intake pattern. With increased knowledge about the extent of dietary supplement use among men with PCa, clinicians will be able to be more observant and ask their patients about such use. A more detailed assessment of what types of dietary supplements are most commonly used by men with PCa may be important for possibilities to give proper advice about supplement use and to deliver prescribed therapies without risking toxic effects or adverse interactions. Also, with knowledge about existing supplement use patterns and the evident interest in diet among men with PCa, we may increase the efficiency of scientific evaluations of dietary interventions to slow prostate tumor growth.
Annals of Oncology
Annals of Oncology 22: 967–972, 2011 doi:10.1093/annonc/mdq456 Published online 6 October 2010
Dietary supplement use patterns in men with prostate cancer: the Cancer Prostate Sweden Study A. Westerlund1, G. Steineck1,2, K. Ba¨lter3, P. Stattin4, H. Gro¨nberg3 & M. Hedelin1,2* 1
Division of Clinical Cancer Epidemiology, Department of Oncology, Institute of Clinical Sciences, Sahlgrenska Academy at University of Gothenburg, Gothenburg; Division of Clinical Cancer Epidemiology, Department of Oncology and Pathology, Karolinska Institutet, Stockholm; 3Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm; 4Department of Surgery and Perioperative Sciences, Urology and Andrology, Umea˚ University Hospital, Umea˚, Sweden 2
Received 1 March 2010; revised 13 April 2010; accepted 22 July 2010
controls in Sweden. Age-adjusted binary regression with an identity link was carried out to estimate prevalence differences and corresponding 95% confidence intervals (CIs). Modifying effects of lifestyle- and diet-related factors were explored by statistical assessment of additive interaction. Results: Among men with PCa, 542 individuals (48%) had used supplements, which was a 10% (95% CI: 5.9%– 15%) higher prevalence than among population controls. Among individuals with high intake of fatty fish, vegetables, and phytoestrogens, but low intake of saturated fat, supplement use was 29% (95% CI: 18%–41%) more common in men with PCa than in population controls. We found no evidence of heterogeneity by categories of education, smoking history, body mass index, fiber, fruit, or phytoestrogen intake, treatment, or disease stage. Conclusion: Supplement use is common in Swedish men with PCa, especially among those with a healthy dietary pattern. Key words: complementary therapies, dietary supplements, prostatic neoplasms
introduction The use of complimentary and alternative medicine (CAM) is common among cancer patients, with reported prevalences ranging between 36% and 70% [1–3]. Of the various types of CAM that cancer patients use, dietary supplements represent a considerable fraction [3, 4]. In general, total CAM use (or use of dietary supplements specifically) in cancer patients or longer term survivors is associated with higher educational level and higher income and in some instances with younger or older age, smoking habits, or disease stage [1–5]. Previous studies indicate that supplement use is associated with dietary habits [6, 7]. Most studies of CAM use among men with prostate cancer (PCa) have been conducted in North American men [5, 7–9]. To our knowledge, European studies are few, relatively small and none has compared CAM use among men with PCa and matched population controls [10–12]. We explored whether men with PCa were more likely to use dietary supplements than men without PCa in a population-based study of Swedish men. In addition, we investigated whether supplement use
*Correspondence to: Dr. M. Hedelin, Division of Clinical Cancer Epidemiology, Department of Oncology, Institute of Clinical Sciences, Sahlgrenska University Hospital, SE-413 45 Gothenburg, Sweden. Tel: +46-70-7401858; Fax: +46-31-820114; E-mail: [email protected]
was dependent on factors such as dietary habits, body mass index (BMI), smoking, education, cancer treatment, or disease status.
materials and methods study population Cancer Prostate in Sweden is a population-based case–control study, with enrollment from January 2001 to September 2002 [13]. Men, 35–79 years of age, with pathologically verified adenocarcinoma of the prostate (International Classification of Diseases-10: C61) were reported to four regional cancer registries in Sweden. Population controls were randomly selected from the Swedish Population Registry and frequency matched by age and geographic residence. Men with advanced PCa disease were defined as those with at least one of the following criteria: tumor–node–metastasis stage, = T3/T4, N+, M+; Gleason score = 8–10; or prostate-specific antigen (PSA) level ‡ 100 ng/ml [14]. Men with localized PCa disease were those not meeting any of the above criteria. Of the 1895 men with PCa who were invited to the study, 1499 (79%) agreed to participate by completing the questionnaire. Of 1684 invited control subjects, 1130 (67%) completed the questionnaire [13]. The investigation was
ª The Author 2010. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For permissions, please email: [email protected]
original article
cancer (PCa) and to what extent lifestyle, disease or other factors influence such use.
Patients and methods: We evaluated supplement use in 1127 men with incident PCa and in 900 population
Downloaded from http://annonc.oxfordjournals.org/ at Florida Atlantic University on July 29, 2015
Background: In a European setting, we know little about the use of dietary supplements among men with prostate
original article approved by the Ethics Committees at Karolinska Institutet and at Umea˚ University.
statistical methods Men with PCa who stated that they had used supplements the current year or that they had initiated supplement use the same year as they were diagnosed were defined as current users. The median time between diagnosis and return of the questionnaire to the study center was 5.6 months (5%–95% percentiles: 2–12 months). Population controls who stated that they had used dietary supplements the current year or that they had initiated supplement use the same year as the questionnaire was returned to the study center, also were defined as current users. Individuals who stated in the questionnaire that they had used supplements the current year but gave information on use before that during the telephone interviews were defined as former users. In the analyses, only current and never users were included, while former users (n = 602) were omitted. In total, 1127 men with PCa and 900 population controls were included in the analyses. Baseline characteristics of cases and controls were compared using a two-sided t-test for continuous variables and a v2 test for categorical variables. In exploring whether individuals with PCa were more likely to use dietary supplements (current versus never user) than subjects without PCa, we used ageadjusted (with age in 5-year intervals) binary regression with an identity or logarithmic link summarized in terms of prevalence differences and prevalence ratios, respectively, and corresponding 95% confidence intervals (CIs). PCa (yes/no) was considered the exposure and dietary supplement use (current user/never user) the outcome (event).
968 | Westerlund et al.
Supplements were divided into subgroups: antioxidants (e.g. vitamins C and E, selenium, beta-carotene, Oxigard, coenzyme Q10), vitamins and/or minerals (e.g. multivitamins with or without minerals, calcium, folate, magnesium, B vitamins, zinc), omega-3 fatty acids, supplements for symptoms of benign prostatic hyperplasia (e.g. pollen preparations, saw palmetto, and field pumpkin preparations), and other supplements (e.g. algae, garlic preparations, ginseng, ginkgo biloba, and other natural remedies). The subcategories serve only as descriptive data as we could not determine whether individuals who had used more than one type of supplement had done so before or after diagnosis (the current year or before that for controls) or both. The association between dietary supplement use and PCa was stratified by lifestyle- and diet-related categories, including education, smoking, BMI, intake of fatty fish, saturated fat, vegetables, fiber, fruit, and food items rich in phytoestrogens. We also considered possible differences in supplement use between men with PCa who had received treatment (n = 813) and those who had not (n = 314) and between those with localized (n = 628) or advanced disease (n = 448) (unknown status, n = 51). Intake of the various foods was categorized into tertiles or quartiles based on the distribution among the population controls. Based on previous results regarding dietary habits generally considered as healthy [13, 17, 18], we created a variable for increasing ‘health index’, which grouped the men according to intake of fatty fish, vegetables, food items rich in phytoestrogens, and saturated fat. The index had three levels: (i) low intake of fatty fish, food items rich in phytoestrogens, and vegetables and high intake of saturated fat; (ii) intermediary intake; and (iii) high intake of fatty fish and food items rich in phytoestrogens and vegetables and low intake of saturated fat. To explore modifying effects of lifestyle- and diet-related factors, statistical assessment of interaction effects was carried out on the additive effect scales. PCa status was represented by two indicator variables comparing men with PCa against men without PCa, and each lifestyle- and dietrelated factor was categorized into three or four groups (see Table 2). Interaction was assessed by the same product terms under a linear odds model, and all analyses were adjusted for age. Analyses were carried out using the STATA System Software, version 10.1.
results characteristics of study participants Baseline characteristics, intake of macronutrients, and use of supplements among the study participants are presented in Table 1. We found no statistically significant differences between subjects with or without PCa, respectively, with regard to BMI, smoking history, level of education, intake of main groups of macronutrients, or overall energy intake. The mean age among men without PCa was somewhat higher than among men with PCa. With one exception, men with PCa were more likely to use supplements than population controls in each dietary supplement subcategory. There was no statistically significant
Volume 22 | No. 4 | April 2011
Downloaded from http://annonc.oxfordjournals.org/ at Florida Atlantic University on July 29, 2015
assessment of food and dietary supplement intake The questionnaire included questions on education, smoking history, treatment type, weight, height, and a validated foodfrequency questionnaire to measure average intake of foods, beverages, and dietary supplements the past 12 months [15]. The respondents were asked if they had used any vitamins, minerals, or other dietary supplements during this time period. They answered by ‘No, I have never eaten any dietary supplements’; ‘Yes, I have eaten dietary supplements before, but not the current year’; or ‘Yes, I have eaten dietary supplements the current year’. Men who stated that they had used supplements the current year or before that were asked to further specify which type(s) of supplements they had used (e.g. multivitamins with or without minerals, vitamin C, vitamin E, selenium, calcium, or fish oil/omega-3 fatty acids). Respondents also could state ever use of specific supplements, such as Cernitol (pollen extract), Curbicin (extracts from saw palmetto fruit and field pumpkin seeds), Echinacea (purple coneflower), garlic, ginseng, coenzyme Q10, and different trademarks of vitamins and minerals. If information was missing from the mailed questionnaire, participants were contacted by phone to complete the missing information. Questionnaire data about the intake of food items were converted into intake of energy and nutrients by linkage to the nutrient database created by the Swedish National Food Administration [16].
Annals of Oncology
original article
Annals of Oncology
Table 1. Selected characteristics of participants with questionnaire data in the Cancer Prostate Sweden Study Characteristic
Men with PCa, n = 1127
Men without PCa, n = 900
0.01 0.46
0.07
0.54
– 0.1
0.1 0.2 0.1
<0.01
0.01
<0.01
<0.01
0.2
0.01
BMI, body mass index; PCa, prostate cancer.
difference in use of omega-3 fatty acid supplements between men with and without PCa (data not shown).
PCa status and use of dietary supplements Five hundred and forty-two men (48%) with PCa and 341 men (38%) without PCa had used any supplement(s) the previous year (Table 2). The use of supplements was 9.9% (95% CI: 5.6%–14.2%) more common in men with PCa than in population controls. Table 2 also shows the prevalence
Volume 22 | No. 4 | April 2011
discussion In this study, we found that dietary supplement use was more common among men with PCa than in population controls. In the group that had the healthiest dietary intake pattern (high intake of fatty fish, vegetables, and food items rich in phytoestrogens and low intake of saturated fat), a higher proportion of men with PCa were supplement users compared with population controls. A handful of studies of CAM use among men with PCa report the prevalence of dietary supplement use specifically. Accordingly, 15%–56% of men with PCa use dietary supplements [3, 5, 8, 9, 19]. In a study that focused exclusively on supplement use among men with PCa, the user prevalence was 73% [7]. Possible explanations for the differences in prevalence rates compared with our data include differing definitions of dietary supplements and differing points in time
doi:10.1093/annonc/mdq456 | 969
Downloaded from http://annonc.oxfordjournals.org/ at Florida Atlantic University on July 29, 2015
Age, mean 6 SD (years) 67 6 7.3 68 6 7.5 BMI, mean 6 SD (kg/m2) 26 6 3.4 26 6 3.4 Education, n (%) Compulsory school, 0–9 532 (47) 423 (47) years Upper secondary, 10–12 431 (38) 372 (41) years University, ‡13 years 161 (14) 99 (11) Missing 3 (1) 6 (1) Smoking history, n (%) Never 444 (39) 341 (38) Ever 667 (59) 542 (60) Missing 16 (2) 17 (2) Treatment, n (%) Hormones 493 (44) – Surgery 248 (22) – Radiotherapy 72 (6) – No treatment 214 (28) 900 Total energy intake, 9245 8913 median (kJ) Proportion of energy intake, median (%) from: Fat (%) 32 32 Protein (%) 16 16 Carbohydrate (%) 50 49 Use of any dietary supplement, n (%) Never 585 (52) 559 (62) Current 542 (48) 341 (38) Use of antioxidants, n (%) Never 840 (75) 715 (79) Ever 287 (25) 185 (21) Use of vitamins and/or minerals, n (%) Never 731 (65) 667 (74) Ever 396 (35) 233 (26) Use of supplements for benign prostatic hyperplasia, n (%) Never 998 (89) 849 (94) Ever 129 (11) 51 (6) Use of omega-3 fatty acids, n (%) Never 998 (89) 814 (90) Ever 129 (11) 86 (10) Use of other supplements, n (%) Never 850 (75) 724 (80) Ever 277 (25) 176 (20)
P value
differences and prevalence ratios of dietary supplement use comparing PCa status, stratified by dietary intakes, BMI, education, and smoking history. When carrying out interaction analysis, we found a significant interaction between PCa status and saturated fat intake (P = 0.01), vegetable intake (P = 0.04), and healthy dietary intake pattern (health index) (P = 0.002) and borderline significance for fatty fish intake (P = 0.06) but not for level of education, smoking history, BMI, or fiber, fruit, or phytoestrogen intake (Table 2). In the group with the healthiest dietary intake pattern, a higher proportion of men with PCa was supplement users compared with population controls (prevalence difference: 29%, 95% CI: 18%–41%). Regardless of PCa status, we found that the proportion of supplement users was higher among men with high educational level (‡10 years) compared with men with low educational level (0–9 years) (prevalence difference: 19, 95% CI: 15%–24%) (data not shown). But we found no evidence of heterogeneity by categories of education (P = 0.86) (Table 2). The use of supplements was more common among younger men than older men, irrespective of PCa status. The difference in prevalence of supplement use comparing men aged 40–59 years with men aged 69–74 years was 13% (95% CI: 5.3%–21%) (data not shown). Furthermore, men who were overweight or obese (BMI > 25) used supplements to a 7% lower degree than normal-weight (BMI < 25) men (95% CI: 212% to 23%) (data not shown). There was no significant difference in rate of overall supplement use between men who had received some kind of treatment (hormones, radiotherapy, or surgery) and those who had not received any treatment (prevalence difference: 3.6, 95% CI: 22.9% to 10%) (data not shown). When specific supplement use was stratified by type of treatment, men who had received hormone therapy were more likely to use omega-3 supplements compared with those who had not received any treatment (prevalence difference: 6.2, 95% CI: 1.8%–11%) (data not shown). Supplement user rates did not significantly differ between men with advanced compared with localized disease (prevalence difference: 22.5, 95% CI: 28.7% to 3.6%) (data not shown). Adjustment for potential confounders (education level, smoking history, etc.) did not substantially change any of the results (data not shown).
original article
Annals of Oncology
Table 2. Prevalence differences and prevalence ratios with corresponding 95% CI for dietary supplement use by levels of PCa status, stratified by education, smoking, BMI, and dietary intake Characteristic
PCa statusa
Dietary supplementb use, n (%) Never Current
All participants
No Yes
559 (62) 585 (52)
341 (38) 542 (48)
0 (ref.) 9.9 (5.6–14.2)
1 (ref.) 1.3 (1.1–1.4)
306 329 207 205 41 49
(72) (62) (56) (48) (41) (30)
117 203 165 226 58 112
(28) (38) (44) (52) (59) (70)
0 9.5 0 8.2 0 11
(ref.) (3.6–15) (ref.) (1.3–15) (ref.) (21.2 to 23)
1 1.4 1 1.2 1 1.2
(ref.) (1.1–1.6) (ref.) (1.0–1.4) (ref.) (0.97–1.4)
0.86
211 221 338 356
(62) (50) (38) (53)
130 223 130 223
(38) (50) (38) (50)
0 12 0 9.1
(ref.) (4.9–19) (ref.) (3.5–15)
1 1.3 1 1.2
(ref.) (1.1–1.5) (ref.) (1.1–1.4)
0.47
186 205 272 278 76 77
(58) (47) (63) (54) (64) (58)
133 235 159 236 42 56
(42) (53) (37) (46) (36) (42)
0 11 0 8.7 0 6.7
(ref.) (4.1–18) (ref.) (2.4–15) (ref.) (25.3 to 19)
1 1.3 1 1.2 1 1.2
(ref.) (1.1–1.5) (ref.) (1.1–1.4) (ref.) (0.88–1.6)
0.22
32 55 467 489 60 41
(60) (60) (62) (53) (61) (39)
21 37 281 440 39 65
(40) (40) (38) (47) (39) (61)
0 0.43 0 9.5 0 21
(ref.) (216 to 17) (ref.) (4.8–14) (ref.) (7.1–35)
1 1.1 1 1.3 1 1.5
(ref.) (0.72–1.6) (ref.) (1.1–1.4) (ref.) (1.1–2.0)
0.06
159 118 134 150 132 173 134 144
(70) (47) (60) (55) (59) (50) (60) (55)
68 131 91 125 92 170 90 116
(30) (53) (40) (45) (41) (50) (40) (45)
0 22 0 4.8 0 7.8 0 5.1
(ref.) (14–31) (ref.) (23.9 to 14) (ref.) (20.40 to 16) (ref.) (23.7 to 14)
1 1.7 1 1.1 1 1.2 1 1.1
(ref.) (1.4–2.2) (ref.) (0.91–1.4) (ref.) (1.0–1.4) (ref.) (0.90–1.4)
0.01
131 131 136 141 138 146 154 161
(60) (55) (60) (53) (61) (52) (68) (49)
88 107 90 125 90 137 73 167
(40) (45) (40) (47) (39) (48) (32) (51)
0 4.5 0 7.4 0 9.0 0 19
(ref.) (24.4 to 14) (ref.) (21.3 to 16) (ref.) (0.42–18) (ref.) (10–27)
1 1.1 1 1.2 1 1.2 1 1.6
(ref.) (0.89–1.3) (ref.) (0.99–1.5) (ref.) (1.0–1.5) (ref.) (1.3–2.0)
0.04
166 183 150 138 126 137 117 127
(74) (61) (67) (61) (56) (50) (52) (39)
59 115 75 90 99 139 108 198
(26) (39) (33) (39) (44) (50) (48) (61)
0 11 0 4.5 0 7.1 0 13
(ref.) (3.3–19) (ref.) (24.2 to 13) (ref.) (21.6 to 16) (ref.) (4.7–22)
1 1.4 1 1.1 1 1.2 1 1.3
(ref.) (1.1–1.9) (ref.) (0.84–1.4) (ref.) (0.98–1.4) (ref.) (1.1–1.5)
0.93
970 | Westerlund et al.
Prevalence ratioc (95% CI)
P value for homogeneity
Volume 22 | No. 4 | April 2011
Downloaded from http://annonc.oxfordjournals.org/ at Florida Atlantic University on July 29, 2015
Education level (years) 0–9 No Yes 10–12 No Yes >13 No Yes Smoking history Never No Yes Ever No Yes BMI <25 No Yes 25 to >30 No Yes ‡30 No Yes Fatty fish intake Never No Yes £2 times/week No Yes >3 times/week No Yes Saturated fat intake (g/d) 0–26 No Yes 26–33 No Yes 33–44 No Yes 44–154 No Yes Vegetable intake (g/d) 0–184 No Yes 185–245 No Yes 245–327 No Yes 327–1768 No Yes Fiber intake (g/d) 0–18 No Yes 18–23 No Yes 23–29 No Yes 29–76 No Yes
Prevalence differencec (95% CI)
original article
Annals of Oncology
Table 2. (Continued) Characteristic
PCa statusa
Fruit intake (g/d) 0–74
Prevalence differencec (95% CI)
Prevalence ratioc (95% CI)
P value for homogeneity
139 126 137 147 125 146 158 160
(61) (51) (61) (53) (57) (53) (69) (51)
88 122 87 128 94 131 72 155
(39) (49) (39) (47) (43) (47) (31) (49)
0 10.4 0 7.3 0 5.8 0 17
(ref.) (1.5–19) (ref.) (21.4 to 16) (ref.) (23.0 to 15) (ref.) (8.8–25)
1 1.3 1 1.2 1 1.1 1 1.6
(ref.) (1.0–1.6) (ref.) (0.96–1.5) (ref.) (0.91–1.3) (ref.) (1.3–2.0)
0.36
142 174 130 132 152 140 135 133
(59) (54) (61) (54) (67) (53) (61) (47)
98 148 83 114 75 126 85 148
(41) (46) (39) (46) (33) (47) (39) (53)
0 4.9 0 7.3 0 14 0 14
(ref.) (23.4 to 13) (ref.) (21.8 to 16) (ref.) (5.2–22) (ref.) (4.8–22)
1 1.1 1 1.2 1 1.4 1 1.4
(ref.) (0.93–1.4) (ref.) (0.96–1.5) (ref.) (1.1–1.8) (ref.) (1.1–1.7)
0.58
84 109 401 424 74 52
(64) (60) (62) (54) (62) (33)
48 72 248 362 45 108
(36) (40) (38) (46) (38) (68)
0 2.9 0 7.8 0 29
(ref.) (28.0 to 14) (ref.) (2.7–13) (ref.) (18–41)
1 1.1 1 1.2 1 1.8
(ref.) (0.83–1.5) (ref.) (1.1–1.4) (ref.) (1.4–2.3)
0.002
a
Men with pathologically verified adenocarcinoma of the prostate (International Classification of Diseases-10:C61) are defined as ‘yes’ and population controls as ‘no’. b Dietary supplement use refers to use after PCa diagnosis (for men with PCa) or the current year (for population controls). Dietary supplements include, e.g. multivitamins, single vitamins and minerals, omega-3 fatty acids, and natural remedies. c Adjusted for age. d Health index: (i) low intake of fatty fish, food items rich in phytoestrogens and vegetables, high intake of saturated fat; (ii) intermediary intake; and (iii) high intake of fatty fish and food items rich in phytoestrogens and vegetables and low intake of saturated fat. BMI, body mass index; CI, confidence interval; PCa, prostate cancer; SD, standard deviation.
in relation to PCa diagnosis when use was measured, as well as varying places of residence of the participants. Regardless of PCa status, we found that men with higher educational level were more likely to use supplements than men without high school education and that younger men were more likely to use supplements than older men. These findings are in line with several studies of CAM use among survivors of PCa as well as other cancers [3, 5–7, 9, 10, 19] and in the general population [20]. Furthermore, our observation that a healthy dietary intake pattern influences the use of supplements finds support in previous investigations. For example, PCa and other cancer survivors with high intake of fruits and vegetables or of fruit, whole grain, and oil had a higher use of supplements [6, 7]. Comparable results were found in a healthy population [20]. Overall dietary supplement use did not differ between men with PCa who had received treatment and those who had not. This is consistent with a smaller study in which no difference in CAM use was found between treatment groups [21]. Wilkinson et al. [8] and Boon et al. [19] did, though, find that men who chose hormonal therapy were more likely to use CAM
Volume 22 | No. 4 | April 2011
compared with those who chose surgery or radiation and CAM nonusers, respectively. We found that men who had received hormonal therapy were more prone to use omega-3 fatty acid supplements compared with men who had not received any treatment. Wiygul et al. [7] found that men who chose other treatment options than surgery or radiation were more prone to use herbals, while in the study by Diefenbach et al. [9], watchful waiting was the only treatment option associated with use of herbals. In terms of disease stage, we found no difference in supplement use between men with localized and advanced disease. Advanced [19] and progressive [8] disease have previously been positively associated with CAM use. Strengths of our study include its population-based design, large size, and complete and rapid case ascertainment. The ethnic homogeneity of our study population reduces the risk of confounding by population stratification. The Swedish cancer registries record almost 100% of all incident cases and since there was limited PSA testing in our study population at the time of enrollment [13], our results pertain mainly to nonPSA-detected clinically significant PCa. Any problem with
doi:10.1093/annonc/mdq456 | 971
Downloaded from http://annonc.oxfordjournals.org/ at Florida Atlantic University on July 29, 2015
No Yes 74–135 No Yes 135–220 No Yes 220–1430 No Yes Phytoestrogen intake (g/d) 0–8 No Yes 9–23 No Yes 23–41 No Yes 41–596 No Yes Health indexd 1 No Yes 2 No Yes 3 No Yes
Dietary supplementb use, n (%) Never Current
original article
funding Swedish Cancer Society (050158 to HG, 080062 to MH); Region Va¨stra Go¨taland, Sweden.
acknowledgements The authors would like to thank all study participants in the CAPS study; Ulrika Lund Unde´n for skillfully coordinating the study center at Karolinska Institutet; Lotta Spa˚ngberg, Berit Andersson, and Britt Eriksson for conducting thorough interviews; and all urologists whose patients were included in the CAPS study.
972 | Westerlund et al.
disclosure The authors declare no conflicts of interest.
references 1. Molassiotis A, Fernadez-Ortega P, Pud D et al. Use of complementary and alternative medicine in cancer patients: a European survey. Ann Oncol 2005; 16: 655–663. 2. Paltiel O, Avitzour M, Peretz T et al. Determinants of the use of complementary therapies by patients with cancer. J Clin Oncol 2001; 19: 2439–2448. 3. Patterson RE, Neuhouser ML, Hedderson MM et al. Types of alternative medicine used by patients with breast, colon, or prostate cancer: predictors, motives, and costs. J Altern Complement Med 2002; 8: 477–485. 4. Velicer CM, Ulrich CM, Velicer CM, Ulrich CM. Vitamin and mineral supplement use among US adults after cancer diagnosis: a systematic review. J Clin Oncol 2008; 26: 665–673. 5. Chan JM, Elkin EP, Silva SJ et al. Total and specific complementary and alternative medicine use in a large cohort of men with prostate cancer. Urology 2005; 66: 1223–1228. 6. Miller P, Demark-Wahnefried W, Snyder DC et al. Dietary supplement use among elderly, long-term cancer survivors. J Cancer Surviv 2008; 2: 138–148. 7. Wiygul JB, Evans BR, Peterson BL et al. Supplement use among men with prostate cancer. Urology 2005; 66: 161–166. 8. Wilkinson S, Gomella LG, Smith JA et al. Attitudes and use of complementary medicine in men with prostate cancer. J Urol 2002; 168: 2505–2509. 9. Diefenbach MA, Hamrick N, Uzzo R et al. Clinical, demographic and psychosocial correlates of complementary and alternative medicine use by men diagnosed with localized prostate cancer. J Urol 2003; 170: 166–169. 10. Wilkinson S, Farrelly S, Low J et al. The use of complementary therapy by men with prostate cancer in the UK. Eur J Cancer Care 2008; 17: 492–499. 11. Ponholzer A, Struhal G, Madersbacher S. Frequent use of complementary medicine by prostate cancer patients. Eur Urol 2003; 43: 604–608. 12. Salmenpera¨ L. The use of complementary therapies among breast and prostate cancer patients in Finland. Eur J Cancer Care 2002; 11: 44–50. 13. Hedelin M, Klint A, Chang ET et al. Dietary phytoestrogen, serum enterolactone and risk of prostate cancer: the cancer prostate Sweden study (Sweden). Cancer Causes Control 2006; 17: 169–180. 14. Sobin LH, Wittekind C (eds): TNM Classification of Malignant Tumours (UICC), 6th edition. New York: Wiley-Liss 2002. 15. Chang ET, Smedby KE, Zhang SM et al. Dietary factors and risk of non-hodgkin lymphoma in men and women. Cancer Epidemiol Biomarkers Prev 2005; 14: 512–520. 16. The National Food Administration’s Food Database. http://www7.slv.se/ Naringssok/SokLivsmedel.aspx (12 August 2010, date last accessed). 17. Hedelin M, Chang ET, Wiklund F et al. Association of frequent consumption of fatty fish with prostate cancer risk is modified by COX-2 polymorphism. Int J Cancer 2007; 120: 398–405. 18. Jakobsen MU, O’Reilly EJ, Heitmann BL et al. Major types of dietary fat and risk of coronary heart disease: a pooled analysis of 11 cohort studies. Am J Clin Nutr 2009; 89: 1425–1432. 19. Boon H, Westlake K, Stewart M et al. Use of complementary/alternative medicine by men diagnosed with prostate cancer: prevalence and characteristics. Urology 2003; 62: 849–853. 20. Foote JA, Murphy SP, Wilkens LR et al. Factors associated with dietary supplement use among healthy adults of five ethnicities: the multiethnic cohort study. Am J Epidemiol 2003; 157: 888–897. 21. Lippert MC, McClain R, Boyd JC, Theodorescu D. Alternative medicine use in patients with localized prostate carcinoma treated with curative intent. Cancer 1999; 86: 2642–2648.
Volume 22 | No. 4 | April 2011
Downloaded from http://annonc.oxfordjournals.org/ at Florida Atlantic University on July 29, 2015
misclassification of disease status (in this study considered as the exposure of interest) therefore is unlikely. Prevalences of dietary supplement use may be influenced by erroneous reports by the respondents, measurement errors are unavoidable. Measurement errors may arise, for example, if the cancer patients were more prone to state their supplement use than the controls. They may, for example, have a greater interest in the possible effect of dietary supplementation on the disease. If the cancer patients overreported their use, this could account for the apparent positive association between PCa status and supplement use. On the other hand, some underreporting is also possible among patients who think they should use more supplements than they currently do. However, most participants (1787 of 2027) were contacted by phone to complete missing information from the questionnaire, including information about supplement use. The interviewers were blinded to the respondents’ PCa status. Therefore, we do not believe that this type of potentially differential misclassification has had any significant impact on our results. Finally, the study was not originally designed to measure whether use of dietary supplements had been initiated or changed after diagnosis. Even though we have information on when supplement use started, our definition is not optimal and we probably have some misclassification. Prior supplement use could lead to misclassification of our binary outcome variable, but such misclassification is likely non-differential and would attenuate any true association. In conclusion, use of dietary supplements among Swedish men with PCa is widespread and even higher in men with a healthy dietary intake pattern. With increased knowledge about the extent of dietary supplement use among men with PCa, clinicians will be able to be more observant and ask their patients about such use. A more detailed assessment of what types of dietary supplements are most commonly used by men with PCa may be important for possibilities to give proper advice about supplement use and to deliver prescribed therapies without risking toxic effects or adverse interactions. Also, with knowledge about existing supplement use patterns and the evident interest in diet among men with PCa, we may increase the efficiency of scientific evaluations of dietary interventions to slow prostate tumor growth.
Annals of Oncology