- Email: [email protected]
Waist circumference, waist-hip ratio, body mass index, and prostate cancer risk: Results from the North-American case-control study Prostate Cancer & Environment Study
Urologic Oncology: Seminars and Original Investigations ] (2015) ∎∎∎–∎∎∎
Original article
Waist circumference, waist-hip ratio, body mass index, and prostate cancer risk: Results from the North-American case-control study Prostate Cancer & Environment Study Katharina Boehm, M.D.a,b,*, Maxine Sunb, Alessandro Larcher, M.D.b,c, Audrey Blanc-Lapierred, Jonas Schiffmann, M.D.a, Markus Graefen, M.D.a, José Sosaa, Fred Saad, M.D.e, Marie-Élise Parentd,f, Pierre I. Karakiewicz, M.D.b,e a Martini-Klinik am Universitätsklinikum Hamburg-Eppendorf, Hamburg, Germany Cancer Prognostics and Health Outcomes Unit, University of Montreal Health Center, Montreal, Canada c Division of Oncology, URI, IRCCS Ospedale San Raffaele, Milan, Italy d INRS-Institut Armand-Frappier, Institut national de la recherche scientifique, Université du Québec, Laval, Canada e Department of Urology, University of Montreal Health Center, Montreal, Canada f Department of Social and Preventive Medicine, University of Montreal, Canada b
Received 30 April 2015; received in revised form 22 June 2015; accepted 10 July 2015
Abstract Introduction: The evidence on the association between anthropometric measures quantifying body fatness and prostate cancer (PCa) risk is not entirely consistent. Associations among waist circumference (WC), waist-hip ratio, body mass index (BMI), and PCa risk were assessed in a population-based case-control study. Patients and methods: The study included 1933 incident PCa cases diagnosed between 2005 and 2009. Population controls were 1994 age-matched (⫾5 y) Montreal residents selected from electoral lists. Information on sociodemographics, medical history including PCa screening, height, weight, and waist and hip circumferences was collected through interviews. Logistic regression was used to assess odds ratios (ORs) for the association between anthropometric measures, and overall and grade-specific PCa. Results: After adjustment for BMI, an excess risk of high-grade PCa (Gleason Z7) was associated with a WC Z102 cm (OR ¼ 1.47 [1.22–1.78]) and with a waist-hip ratio 41.0 (OR ¼ 1.20 [1.01–1.43]). Men with a BMI Z30 kg/m2 had a lower risk of PCa, regardless of grade. Restricting to subjects recently screened for PCa did not alter findings. Conclusion: Elevated BMI was associated with a lower risk of PCa, regardless of grade. Contrastingly, abdominal obesity, when adjusted for BMI, yielded results in the opposite direction. Taken together, our observations suggest that the specific body fat distribution (abdominal), for a given BMI, is a predictor of PCa risk, whereas BMI alone is not. BMI and abdominal obesity, especially when measured by the WC, should be examined conjointly in future studies on this issue and may require consideration at patient counseling. r 2015 Elsevier Inc. All rights reserved. Keywords: Prostate cancer; Obesity; Waist circumference; Waist-hip ratio; Body mass index; Abdominal fat; Case-control study
Disclosures: This study was supported financially through grants from the Canadian Cancer Society Canada, the Cancer Research Society, the Fonds de la recherche du Québec—Santé (FRQS), FRQS-RRSE, and the Ministère du Développement économique, de l’Innovation et de l’Exportation du Québec. Marie-Élise Parent and Pierre Karakiewicz held career awards from the FRQS. Fred Saad holds the University of Montreal Endowed Chair in Prostate Cancer Research. * Corresponding author. Tel.: þ1-514-890-8000x35335; fax: þ1-514227-5103. E-mail addresses: [email protected], katharina. [email protected] (K. Boehm). http://dx.doi.org/10.1016/j.urolonc.2015.07.006 1078-1439/r 2015 Elsevier Inc. All rights reserved.
1. Introduction Obesity and overweightness are on sharp rise in western countries. In 2009 to 2010, 77% of North-American men older than 60 years were classified as overweight (body mass index [BMI]: 25–29.9 kg/m2) or obese (BMI Z 30 kg/m2) [1]. Obesity is associated with an increased risk of several cancers, such as colorectal and breast cancers [2]. Anthropometric measures of body fatness include BMI, waist circumference (WC), and waist-hip ratio (WHR).
2
K. Boehm et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–7
However, inconsistent results have been reported with respect to the association between obesity and PCa risk [3–6]. Specifically, in a large case-control study from Europe [7], increased BMI was inversely related to low-grade PCa risk and no relationship was found between BMI and high-grade PCa [7]. Similar results were also found in a prospective cohort study [8]. The same study showed results in the opposite direction when obesity and high-grade PCa were examined, similar to those of other authors [9]. Conversely, a European prospective cohort study showed no association between elevated BMI and risk of PCa, regardless of grade [10]. The lack of consensus regarding the magnitude and direction of the association between BMI and PCa risk is still debatable. In consequence, it is possible that the relationship between BMI and PCa varies from one population to another or that BMI alone is a poor indicator for obesity and other anthropometric measures alone, or in combination with BMI, might better illustrate body fat distribution. Several hypotheses have been proposed to explain an inverse relationship between BMI and PCa risk. First, it may ensue from lower PCa detection rates among men with elevated BMI, due to a dilutional effect of PSA in men with BMI elevation [11–15]. Second, it is also known that the metabolic syndrome, of which abdominal obesity is an integral part, is associated with a lower risk of PCa diagnosis [16,17]. Third, obesity may also be correlated with a low physical activity, which is possibly associated with PCa [18]. However, the latest evidence linking anthropometric measures and PCa risk was recently reviewed by the World Cancer Research Fund International organization [3]. Although no conclusion could be drawn for total or nonadvanced PCa, recent data were suggestive of a dose-response trend between each of these indicators and the risk of advanced PCa. Based on this, we examined the association between various anthropometric measures of body fatness and PCa risk in a study population constituted of men from Montreal, Canada.
2. Patients and methods 2.1. Study population We relied on data from the population-based case-control study Prostate Cancer & Environment Study (PROtEuS), which recruited subjects in Montreal, Canada, between 2005 and 2012. The study design was previously described [19–21]. In brief, the study population consisted of residents of Montreal metropolitan area, aged o76 years at diagnosis or recruitment. Cases consisted of histologically confirmed, newly diagnosed PCa and were actively ascertained through pathology departments across 7 hospitals in Montreal between 2005 and 2009, covering 480% of all cases diagnosed in the catchment area. Controls had no PCa diagnosis at the time of interview. Age-matched (5-y intervals) controls from the same residential area were randomly selected. Participation rates
were 79.4% for cases and 55.5% for controls. In analyses comparing sociodemographic characteristics of participants and nonparticipants, little differences emerged. Committees of all participating institutions approved the study protocol and all subjects provided informed consent. 2.2. Measures Subjects provided information about sociodemographic characteristics, lifestyle, and their medical history, including the frequency of PSA measurements and digital rectal examinations (DRE), physical activity (not very active, moderately active, and very active), and presence of diabetes (yes/no) by interviews. Anthropometric factors included selfreported current height and weight 2 years before the index date (i.e., diagnosis or interview), referred to as recent BMI. The interviewer measured waist (1 in above the umbilicus) and hip circumferences (maximum) in a standardized fashion. Missing values were recorded in 323 and 328 patients for WC and WHR, respectively. For cases, information about Gleason scores and PSA levels at diagnosis was extracted from patient's files and pathology reports. 2.3. Statistical analyses Our analyses focused on 3 anthropometric indicators of body fatness as potential risk factors: (1) WC, (2) WHR, and (3) recent BMI. WHR was calculated by dividing the waist by the hip circumference. For WC, a cutoff of 102 cm was used to define abdominal obesity, as recommended by the World Health Organization [22]. The cutoff for analyses on WHR was set at 1.0, according to the distribution among controls (mean ¼ 0.99 and median ¼ 1.0). Sensitivity analyses were performed, using 2 different cutoffs (0.97 and 1.02), based on tertiles among controls. BMI was calculated as recent weight divided by the square of the height. BMI was categorized according to the definition of the World Health Organization, into underweight (BMI o 18.5), normal weight (BMI: 18.5–24.9), overweight (BMI: 25.0–29.9), obese class I (BMI: 30.0– 34.9), obese class II (BMI: 35.0–39.9), and extreme obesity (BMI 4 40). For analyses, we combined the underweight and normal weight categories, as well as the obese class I, class II, and extreme obesity categories. Unconditional logistic regression was used to test associations between the 3 anthropometric measures and PCa risk. All models included age at index date, PCa family history in first-degree relatives, ancestry (European, Black, Asian, other, and unknown), physician visits per year (o1, 1–3, 43, and unknown), and number of PSA measurements within the 5 years preceding PCa diagnosis/interview (r4 PSA tests, 44 PSA tests, and unknown). Multivariate logistic regression was used to assess associations with overall PCa risk. Multinomial logistic regression was applied for analyses considering low-grade (Gleason r6) or high-grade (Gleason Z7) PCa. Additional
K. Boehm et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–7
3
analyses tested the association between WC (Z102 cm) and PCa risk, after adjustment for BMI (continuous) and among subgroups stratified according to BMI categories (BMI o 25, BMI: 25.0–29.9, and BMI Z 30). The same approach was used to test the association with WHR (41 vs. r1), adjusting for and stratifying by BMI. Finally, to assess the possibility of a potential PCa detection bias, subgroup analyses were conducted restricting to individuals with PSA testing and/or DRE within 2 years before the index date. Further analyses on BMI, WC, and WHR were adjusted for diabetes and physical activity. All statistical analyses were performed using R version 3.1.0 (R Project for Statistical Computing, www.R-project.org). A 2-sided P o 0.05 was considered to be statistically significant.
Table 1 Characteristics of cases (n ¼ 1,921) and controls (n ¼ 1,982) participating in PROtEuS, Montreal, Canada, 2005–2012
3. Results
PCa family historya, n (%) No 1,728 (87.2) Yes 198 (10) Don't know 56 (2.8)
1,411 (73.5) 448 (23.3) 62 (3.2)
o0.01
Ancestry, n (%) European African Asian Other Don't know
The study consisted of 1933 PCa cases and 1994 controls. After exclusion of 24 patients with unknown BMI, 1921 PCa cases and 1982 controls remained for analyses. Cases were slightly younger than controls (Table 1). Cases were more likely than controls to have a first-degree family history of PCa and to be of African ancestry. The median number of PSA measurements within the 5 years before the index date was similar between cases and controls. The interquartile range (IQR) for WC was 88 to 107 cm (median ¼ 98 cm). The IQR for WHR was 0.65 to 1.34 (mean ¼ 0.99, median ¼ 1.0). Regarding recent BMI, expressed in kg/m2, an IQR of 24 to 29 (mean ¼ 27.0, median ¼ 26.5) was recorded (Table 1). 3.1. Association between WC and PCa Abdominal obesity, as defined by a WC Z 102 cm, was not significantly associated with overall PCa (odds ratio [OR] ¼ 1.03 [0.89–1.19]), low-grade PCa (OR ¼ 0.86 [0.71–1.03]), or high-grade PCa (OR ¼ 1.17 [0.99–1.38]) (Table 2). When adjusted for BMI, elevated WC was associated with an increased risk of overall PCa (OR ¼ 1.23 [1.05–1.46]), with the association more pronounced for high-grade PCa (OR ¼ 1.47 [1.22–1.78]) (Table 3). In analyses stratified according to BMI categories, a positive association between WC Z102 cm and high-grade PCa was observed among overweight (OR 1.41 [1.11–1.78]) as well as obese (OR 1.77 [1.11–2.82]) men. There was no statistically significant interaction between WC and categorical BMI on PCa risk (Poverweight * WC ¼ 0.8; Pobese * WC ¼ 0.9). 3.2. Association between WHR and PCa No association was found between WHR 4 1.0 and overall, as well as grade-specific, PCa risk (Table 2). Similarly, sensitivity analyses suggested no association between WHR 4 0.97 and 41.02, and PCa (data not shown). In
Age, median (IQR) 2
Controls
Cases
n ¼ 1,982
n ¼ 1,921
65 (61–70)
P value
64 (59–69)
o0.01
Recent BMI, kg/m ; median (IQR)
26.6 (24.3–29.5)
Waist circumference, cm; median (IQR)
98.1 (88–108)
99.0 (89–107)
0.9
Hip circumference, cm; median (IQR)
99.2 (89–107)
99.2 (90–107)
0.7
Waist-hip ratio, median (range)
26.8, 26.3 (24.1–29.1)
1.0 (0.65–1.29)
1.0 (0.81–1.34)
0.006
0.5
1,680 86 71 131 14
(84.8) (4.3) (3.6) (6.6) (0.7)
1,687 124 24 74 12
(87.8) (6.5) (1.2) (3.9) (0.6)
o0.01
Family income, n (%) o20,000$ 20,000–29,999$ 30,000–49,999$ 50,000–79,999$ 80,000 and more Other
240 248 461 409 428 196
(12.1) (12.5) (23.3) (20.6) (21.6) (9.9)
220 262 446 423 425 145
(11.5) (13.6) (23.2) (22) (22.1) (7.5)
0.1
Education, n (%) Elementary High School College University Other
423 576 374 608 1
(21.3) (29.1) (18.9) (30.7) (0.1)
444 576 311 588 2
(23.1) (30) (16.2) (30.6) (0.1)
0.2
347 1,328 242 4
(18.1) (69.1) (12.6) (0.2)
0.03
Annual physician visits 5 years ago, n (%) oonce 385 (19.4) 1–3 1,389 (70.1) 43 208 (10.5) Unknown 0 (0) No of PSA 4.1, 5.0 (2.0–5.0) measurements,b mean, median (IQR) a
4.4, 5.0 (2.0–5.0) o0.01
First-degree relative with PCa. Within the last 5 years before the index date.
b
analyses of WHR 4 1.0 adjusted for BMI, a weak albeit statistically significant association for high-grade PCa risk was recorded (OR ¼ 1.20 [1.01–1.43]) (Table 4). In subgroup analyses according to BMI categories, no significant association between WHR and PCa risk was observed (Table 4). No interaction was found between categories of WHR and BMI, and PCa risk (Poverweight * WHR ¼ 0.3; Pobese * WHR ¼ 0.2).
4
Controls, n ¼ 1,982
Overall PCa
Low-grade PCab
High-grade PCac
Restricted to participants who have been screenedd
Cases, n ¼ 1,921
OR (95% CI)
Cases, n ¼ 807
OR (95% CI)
Cases, n ¼ 1,114
OR (95% CI)
Controls, n ¼ 1,505
Cases, n ¼ 1,903
OR (95% CI)
Waist circumference, cme o102 1,073 (54.1) Z102 711 (35.9)
1,074 (55.9) 722 (37.6)
1.00 (Ref.) 1.03 (0.89–1.19)
485 (60.1) 273 (33.8)
1.00 (Ref.) 0.86 (0.71–1.03)
589 (52.9) 449 (40.3)
1.00 (Ref.) 1.17 (0.99–1.38)
816 (41.2) 551 (27.8)
1,066 (55.5) 717 (37.3)
1.00 (Ref.) 1.03 (0.89–1.20)
WHRf r1.0 41.0
1,077 (54.3) 705 (35.6)
1,090 (56.7) 703 (36.6)
1.00 (Ref.) 1.03 (0.89–1.19)
474 (58.7) 284 (35.2)
1.00 (Ref.) 0.95 (0.79–1.14)
616 (55.3) 419 (37.6)
1.00 (Ref.) 1.09 (0.92–1.28)
817 (41.2) 550 (27.7)
1,084 (56.4) 696 (36.2)
1.00 (Ref.) 0.95 (0.82–1.10)
609 (30.7) 940 (47.4) 433 (21.8)
648 (33.7) 922 (48.0) 351 (18.3)
1.00 (Ref.) 0.87 (0.74–1.01) 0.72 (0.60–0.87)**
277 (34.3) 381 (47.2) 149 (18.5)
1.00 (Ref.) 0.83 (0.68–1.00) 0.71 (0.55–0.90)**
371 (33.3) 541 (48.6) 202 (18.1)
1.00 (Ref.) 0.89 (0.75–1.06) 0.73 (0.59–0.91)**
436 (22.0) 744 (37.5) 325 (16.4)
644 (33.5) 905 (47.1) 354 (18.4)
BMI, kg/m2 o25 25–29.9 Z30
1.00 (Ref.) 0.83 (0.71–0.98)* 0.71 (0.58–0.87)**
a Models were adjusted for age, ancestry, first-degree family history of prostate cancer, annual physician visits, and number of PSA tests within 5 years before index date (except subgroup restricted to screened individuals). b Low-grade prostate cancer defined as Gleason grade r6. c High-grade prostate cancer defined as Gleason grade 47. d Screening defined as PSA test and/or DRE within 2 years before index date. e Missing values for 323 patients. f Missing values for 328 patients. * P o 0.05. ** P o 0.01.
K. Boehm et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–7
Table 2 Adjusteda odds ratio (OR) and 95% CI for the association between waist circumference, waist-hip ratio, body mass index, and incident prostate cancer, PROtEuS, Montreal, Canada, 2005–2012
K. Boehm et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–7
5
Table 3 Adjusteda odds ratios (OR) and 95% CI for the association between waist circumference and incident PCa, adjusted and stratified according to body mass index, PROtEuS, Montreal, Canada, 2005–2012 Waist circumference, cm
Adjusted for BMI o102 Z102
Controls, n ¼ 1,784
1,073 (60.1) 711 (39.9)
Overall PCa
Low-grade PCab
High-grade PCac
Cases, n ¼ 1,796
OR (95% CI)
Cases, n ¼ 758
OR (95% CI)
Cases, n ¼ 1,038
OR (95% CI)
1,074 (59.8) 722 (40.2)
1.00 (Ref.) 1.23 (1.05–1.46)
485 (64.0) 273 (36.0)
1.00 (Ref.) 0.97 (0.78–1.20)
589 (56.7) 449 (43.3)
1.00 (Ref.) 1.47 (1.22–1.78)***
532 (29.6) 69 (3.8)
1.00 (Ref.) 1.24 (0.81–1.88)
227 (29.9) 30 (4.0)
1.00 (Ref.) 1.28 (0.77–2.14)
305 (29.4) 39 (3.8)
1.00 (Ref.) 1.20 (0.75–1.93)
470 (26.2) 400 (22.3)
1.00 (Ref.) 1.17 (0.95–1.43)
218 (28.8) 142 (18.7)
1.00 (Ref.) 0.89 (0.69–1.16)
252 (24.3) 253 (24.4)
1.00 (Ref.) 1.41 (1.11–1.78)**
72 (4.0) 258 (14.4)
1.00 (Ref.) 1.27 (0.88–1.83)
40 (5.3) 101 (13.3)
1.00 (Ref.) 0.87 (0.55–1.38)
32 (3.1) 157 (15.1)
1.00 (Ref.) 1.77 (1.11–2.82)*
BMI categories o25 kg/m2 (underweight or normal weight) o102 481 (27.0) Z102 52 (2.9) 25–29.9 kg/m2 (overweight) o102 491 (27.5) Z102 365 (20.5) Z30 kg/m2 (obese) o102 101 (5.7) Z102 294 (16.5)
a Models were adjusted for age, ancestry, first-degree family history of prostate cancer, annual physician visits, and number of PSA tests within 5 years before index date. b Low-grade prostate cancer defined as Gleason grade r6. c High-grade prostate cancer defined as Gleason grade Z7. * P o 0.05. ** P o 0.01. *** P o 0.001.
3.3. Association between recent BMI and PCa Compared with men underweight or of normal weight, overweight men were at lower risk of PCa (OR ¼ 0.87 [0.74–1.01]). Corresponding ORs for low-grade and highgrade PCa were 0.83 (0.68–1.00) and 0.89 (0.75–1.06), respectively (Table 2). The OR for overall PCa was 0.72 (0.60–0.87) for obese men compared to normal weighted men. Obesity was also inversely related to low-grade (OR ¼ 0.71 [0.55–0.90]) and high-grade (OR ¼ 0.73 [0.59–0.91]) PCa (Table 2). In all of the earlier analyses, further adjustments for PSA and DRE testing had minimal effect on results, nor did analyses restricted to individuals with PSA and/or DRE testing within 2 years before the index date, as well as adjustment for diabetes and physical activity.
4. Discussion The Review Panel from the Continuous Update Project of the World Cancer Research Fund International [3] recently concluded that greater body fatness is probably a cause of advanced PCa. This is an important observation as modifiable risk factors for PCa could not yet be identified. Nevertheless, additional evidence based on large, wellconducted studies is needed to be able to ascribe body fatness as a definite cause. The PROtEuS database enabled us to assess associations between several indicators of fat
mass, which can, independently or in combination, be associated with incident PCa. It also provided us with the opportunity to conduct analyses according to PCa grade, as specific relationships may exist according to PCa grade strata [23,24]. No significant association between WC, WHR, and PCa risk was observed in our data before adjustment for BMI. Similar to these, no association between WC or WHR and PCa risk (unadjusted for BMI) was found in the ProtecT cohort [7]. Once BMI was adjusted for in our data, excess risks of high-grade PCa were observed among men with elevated WC and, to a lesser extent, WHR. As the combination of WC or WHR with BMI might better predict PCa risk than when considered independently, we further tested associations for WC and WHR according to 3 frequently used BMI categories, as previously reported [10]. An elevated WC was associated with a higher risk of high-grade PCa in both overweight and obese individuals. Contrastingly, there was no clear evidence of an association between elevated WHR and PCa across BMI categories. Both WC and WHR have been used in the past as indices of central obesity. However, they do not necessarily predict health outcomes equally [25]. WC is an indicator of visceral fat mass, and is closely related to the percentage of fat mass, particularly in men [26]. It may thus be that the WC captures better than WHR the adiposity distribution that relates to PCa risk. Additionally, our results showed that as compared with men underweight or of normal weight based on BMI values, men who were overweight or obese had a lower
K. Boehm et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–7
6
Table 4 Adjusteda odds ratios (OR) and 95% CI for the association between waist-hip ratio and incident prostate cancer, adjusted and stratified according to body mass index, PROtEuS, Montreal, Canada, 2005–2012 Waist-hip ratio
Adjusted for BMI r1.0 41.0
Controls, n ¼ 1,782
1,077 (60.4) 705 (39.6)
Overall PCa
Low-grade PCab
High-grade PCac
Cases n ¼ 1,793
OR (95% CI)
Cases n ¼ 758
OR (95% CI)
Cases n ¼ 1,035
OR (95% CI)
1,090 (60.8) 703 (39.2)
1.00 (Ref.) 1.13 (0.97–1.31)
474 (62.5) 284 (37.5)
1.00 (Ref.) 1.03 (0.85-1.25)
616 (59.5) 419 (40.5)
1.00 (Ref.) 1.20 (1.01–1.43)
(26.1) (7.4)
1.00 (Ref.) 1.33 (0.97–1.84)
199 (26.3) 58 (7.7)
1.00 (Ref.) 1.40 (0.95–2.07)
269 (26.0) 74 (7.2)
1.00 (Ref.) 1.29 (0.90–1.85)
(27.4) (20.8)
1.00 (Ref.) 1.09 (0.89–1.34)
210 (27.7) 150 (19.8)
1.00 (Ref.) 1.00 (0.77–1.30)
281 (27.1) 223 (21.5)
1.00 (Ref.) 1.16 (0.92–1.46)
(7.3) (11.0)
1.00 (Ref.) 0.98 (0.72–1.35)
65 (8.5) 76 (10.0)
1.00 (Ref.) 0.76 (0.51–1.14)
66 (6.4) 122 (11.8)
1.00 (Ref.) 1.21 (0.83–1.76)
BMI categories 1) o25 kg/m2 (underweight or normal weight) r1.0 434 (24.4) 468 41.0 99 (5.6) 132 2) 25–29.9 kg/m2 (overweight) r1.0 492 (27.6) 491 41.0 362 (20.3) 373 3) Z30 kg/m2 (obese) r1.0 151 (8.4) 131 41.0 244 (13.7) 198
a Models were adjusted for age, ancestry, first-degree family history of prostate cancer, annual physician visits, and number of PSA tests within 5 years before index date. b Low-grade prostate cancer defined as Gleason grade r6. c High-grade prostate cancer defined as Gleason grade Z7.
risk of PCa, regardless of grade. Our findings concerning low-grade PCa risk are in line with the results from the Prostate Cancer Prevention Trial (PCPT), a prospective cohort study from the United States of America [8], showing an inverse association between BMI and lowgrade PCa. They are also in agreement with contemporary data provided by a large nested case-control study from the ProtecT (Prostate testing for cancer and Treatment) trial in the United Kingdom, where BMI was associated with a decreased risk of low-grade PCa [7]. The current evidence overall supports a positive relationship between markers of abdominal obesity and risk of advanced PCa [3] and our results support that abdominal obesity is associated with high-grade PCa risk. Some of the differences in part of our results vs. those previously reported may be due to population differences. For example, the proportion of high-grade PCa cases was higher in our study population (13%) than in the SELECT cohort (7%) or in the PROTECT cohort (5%). Patients of the current study were older, on average (64 y) compared with those in the European Prospective Investigation Into Cancer and Nutrition (EPIC) trial (53 y) and SELECT cohort (52 y). Nevertheless, several characteristics of our population were comparable to those in other populations. For example, the proportion of obesity in the current study (20%) was comparable to those reported in the EPIC trial (20%) and the PCPT ( 23%). The EPIC trial originated from 8 European countries and recruited subjects between 1992 and 2000. The PCPT stems from the United States and recruited subjects between 1994 and 2003. Conversely, the SELECT cohort, originating from the United States as well as Puerto Rico and Canada, is more recent (2001–2008),
and it reports a prevalence of obesity of nearly 30%. Our findings may also explain differences in the results of the aforementioned cohort studies. Body fat distribution might be different across populations, even within the same BMI category. For example, in some populations abdominal obesity might most frequently be present in individuals with a high BMI, whereas in other populations abdominal obesity might be present also in men with normal BMI. A possible explanation for this could be dietary pattern, which differs across populations. The present study has important strengths: its large sample size, including some 2000 histologically confirmed incident PCa cases, ability to control for potential confounding factors, especially PCa screening history, and valid and reliable information about anthropometric measurements and assessment of PCa risk according to cancer aggressiveness. Another major advantage of the current study is its population-based design. Our study also has some limitations. Reporting errors based on self-reports cannot be ruled out regarding height and weight, and this would have influenced BMI. However, most validity studies comparing current reported and measured values for weight and height indicate very high correlation coefficients [27,28]. Self-reported values thus appear sufficiently accurate so that their error would have minimal effect on epidemiological measures of association [29]. Nonetheless, other anthropometric measurements, such as weight and hip circumferences, were measured by interviewers following a preestablished protocol. Another limitation relates to response rates, which were relatively high for cases and less for controls. However, in analyses comparing sociodemographic characteristics of participants
K. Boehm et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–7
and nonparticipants, little differences emerged, reassuring against the presence of selection bias in the study. Finally, other unmeasured confounders could be at play in assessing obesity measurements, as well as PCa risk.
[11]
[12]
5. Conclusion Elevated BMI was associated with a lower risk of PCa, regardless of grade. Contrastingly, abdominal obesity, when adjusted for BMI, yielded results in the opposite direction, i. e., an increased risk of high-grade PCa. Taken together, our observations suggest that the specific body fat distribution (abdominal), for a given BMI, is a predictor of PCa risk, whereas BMI alone is not. Moreover, based on our data, WC appears to a better indicator of central obesity than WHR in predicting PCa risk. In consequence, future studies on this topic should consider both abdominal and overall obesity levels. From a clinical perspective, taking into account several anthropometric indicators conjointly such as BMI and WC may lead to a better prediction of PCa risk. References [1] Flegal KM, Carroll MD, Kit BK, Ogden CL. Prevalence of obesity and trends in the distribution of body mass index among US adults, 1999–2010. J Am Med Assoc 2012;307:491–7. [2] Renehan AG, Tyson M, Egger M, Heller RF, Zwahlen M. Body-mass index and incidence of cancer: a systematic review and meta-analysis of prospective observational studies. Lancet 2008;371:569–78. [3] WCRFI. Continuous Update Project Report: Diet, Nutrition, Physical Activity, and Prostate Cancer. Available at: http://www.wcrf.org/ sites/default/files/Prostate-Cancer-2014-Report.pdf 2014, [accessed 04.12.14]. [4] Baillargeon J, Platz EA, Rose DP, et al. Obesity, adipokines, and prostate cancer in a prospective population-based study. Cancer Epidemiol Biomarkers Prev 2006;15:1331–5. [5] Baillargeon J, Rose DP. Obesity, adipokines, and prostate cancer (review). Int J Oncol 2006;28:737–45. [6] Giovannucci E, Rimm EB, Liu Y, et al. Body mass index and risk of prostate cancer in U.S. health professionals. J Natl Cancer Inst 2003; 95:1240–4. [7] Dimitropoulou P, Martin RM, Turner EL, et al. Association of obesity with prostate cancer: a case-control study within the population-based PSA testing phase of the ProtecT study. Br J Cancer. 2011;104:875–81. [8] Gong Z, Neuhouser ML, Goodman PJ, et al. Obesity, diabetes, and risk of prostate cancer: results from the prostate cancer prevention trial. Cancer Epidemiol Biomarkers Prev 2006;15:1977–83. [9] Liang Y, Ketchum NS, Goodman PJ, Klein EA, Thompson IM Jr. Is there a role for body mass index in the assessment of prostate cancer risk on biopsy? J Urol 2014;192:1094–9. [10] Pischon T, Boeing H, Weikert S, et al. Body size and risk of prostate cancer in the European prospective investigation into
[13]
[14]
[15]
[16]
[17]
[18] [19]
[20]
[21]
[22]
[23] [24] [25]
[26]
[27]
[28]
[29]
7
cancer and nutrition. Cancer Epidemiol Biomarkers Prev 2008;17: 3252–61. Rundle A, Richards C, Neugut AI. Body composition, abdominal fat distribution, and prostate-specific antigen test results. Cancer Epidemiol Biomarkers Prev 2009;18:331–6. Rundle A, Neugut AI. Obesity and screening PSA levels among men undergoing an annual physical exam. Prostate 2008;68:373–80. Culp S, Porter M. The effect of obesity and lower serum prostatespecific antigen levels on prostate-cancer screening results in American men. BJU Int 2009;104:1457–61. Banez LL, Hamilton RJ, Partin AW, et al. Obesity-related plasma hemodilution and PSA concentration among men with prostate cancer. J Am Med Assoc 2007;298:2275–80. Baillargeon J, Pollock BH, Kristal AR, et al. The association of body mass index and prostate-specific antigen in a population-based study. Cancer 2005;103:1092–5. Tande AJ, Platz EA, Folsom AR. The metabolic syndrome is associated with reduced risk of prostate cancer. Am J Epidemiol 2006;164:1094–102. De Nunzio C, Aronson W, Freedland SJ, Giovannucci E, Parsons JK. The correlation between metabolic syndrome and prostatic diseases. Eur Urol 2012;61:560–70. Kruk J, Czerniak U. Physical activity and its relation to cancer risk: updating the evidence. Asian Pac J Cancer Prev 2013;14:3993–4003. Parent ME, Goldberg MS, Crouse DL, et al. Traffic-related air pollution and prostate cancer risk: a case-control study in Montreal, Canada. Occup Environ Med 2013;70:511–8. Blanc-Lapierre A, Weiss D, Parent ME. Use of oral anticoagulants and risk of prostate cancer: a population-based case-control study in Montreal, Canada. Cancer Causes Control 2014;25:1159–66. Spence AR, Rousseau MC, Parent ME. Sexual partners, sexually transmitted infections, and prostate cancer risk. Cancer Epidemiol 2014;38:700–7. Nishida C, Ko GT, Kumanyika S. Body fat distribution and noncommunicable diseases in populations: overview of the 2008 WHO Expert Consultation on Waist Circumference and Waist-Hip Ratio. Eur J Clin Nutr 2010;64:2–5. Freedland SJ, Platz EA. Obesity and prostate cancer: making sense out of apparently conflicting data. Epidemiol Rev 2007;29:88–97. Allott EH, Masko EM, Freedland SJ. Obesity and prostate cancer: weighing the evidence. Eur Urol 2013;63:800–9. Cheong KC, Ghazali SM, Hock LK, et al. The discriminative ability of waist circumference, body mass index and waist-to-hip ratio in identifying metabolic syndrome: Variations by age, sex and race. Diabetes Metab Syndr 2015;9:74–8. Bosy-Westphal A, Geisler C, Onur S, et al. Value of body fat mass vs anthropometric obesity indices in the assessment of metabolic risk factors. Int J Obes 2006;30:475–83. Stevens J, Keil JE, Waid LR, Gazes PC. Accuracy of current, 4-year, and 28-year self-reported body weight in an elderly population. Am J Epidemiol 1990;132:1156–63. Casey VA, Dwyer JT, Berkey CS, Coleman KA, Gardner J, Valadian I. Long-term memory of body weight and past weight satisfaction: a longitudinal follow-up study. Am J Clin Nutr 1991;53: 1493–8. Rimm EB, Stampfer MJ, Colditz GA, Chute CG, Litin LB, Willett WC. Validity of self-reported waist and hip circumferences in men and women. Epidemiology 1990;1:466–73.
Original article
Waist circumference, waist-hip ratio, body mass index, and prostate cancer risk: Results from the North-American case-control study Prostate Cancer & Environment Study Katharina Boehm, M.D.a,b,*, Maxine Sunb, Alessandro Larcher, M.D.b,c, Audrey Blanc-Lapierred, Jonas Schiffmann, M.D.a, Markus Graefen, M.D.a, José Sosaa, Fred Saad, M.D.e, Marie-Élise Parentd,f, Pierre I. Karakiewicz, M.D.b,e a Martini-Klinik am Universitätsklinikum Hamburg-Eppendorf, Hamburg, Germany Cancer Prognostics and Health Outcomes Unit, University of Montreal Health Center, Montreal, Canada c Division of Oncology, URI, IRCCS Ospedale San Raffaele, Milan, Italy d INRS-Institut Armand-Frappier, Institut national de la recherche scientifique, Université du Québec, Laval, Canada e Department of Urology, University of Montreal Health Center, Montreal, Canada f Department of Social and Preventive Medicine, University of Montreal, Canada b
Received 30 April 2015; received in revised form 22 June 2015; accepted 10 July 2015
Abstract Introduction: The evidence on the association between anthropometric measures quantifying body fatness and prostate cancer (PCa) risk is not entirely consistent. Associations among waist circumference (WC), waist-hip ratio, body mass index (BMI), and PCa risk were assessed in a population-based case-control study. Patients and methods: The study included 1933 incident PCa cases diagnosed between 2005 and 2009. Population controls were 1994 age-matched (⫾5 y) Montreal residents selected from electoral lists. Information on sociodemographics, medical history including PCa screening, height, weight, and waist and hip circumferences was collected through interviews. Logistic regression was used to assess odds ratios (ORs) for the association between anthropometric measures, and overall and grade-specific PCa. Results: After adjustment for BMI, an excess risk of high-grade PCa (Gleason Z7) was associated with a WC Z102 cm (OR ¼ 1.47 [1.22–1.78]) and with a waist-hip ratio 41.0 (OR ¼ 1.20 [1.01–1.43]). Men with a BMI Z30 kg/m2 had a lower risk of PCa, regardless of grade. Restricting to subjects recently screened for PCa did not alter findings. Conclusion: Elevated BMI was associated with a lower risk of PCa, regardless of grade. Contrastingly, abdominal obesity, when adjusted for BMI, yielded results in the opposite direction. Taken together, our observations suggest that the specific body fat distribution (abdominal), for a given BMI, is a predictor of PCa risk, whereas BMI alone is not. BMI and abdominal obesity, especially when measured by the WC, should be examined conjointly in future studies on this issue and may require consideration at patient counseling. r 2015 Elsevier Inc. All rights reserved. Keywords: Prostate cancer; Obesity; Waist circumference; Waist-hip ratio; Body mass index; Abdominal fat; Case-control study
Disclosures: This study was supported financially through grants from the Canadian Cancer Society Canada, the Cancer Research Society, the Fonds de la recherche du Québec—Santé (FRQS), FRQS-RRSE, and the Ministère du Développement économique, de l’Innovation et de l’Exportation du Québec. Marie-Élise Parent and Pierre Karakiewicz held career awards from the FRQS. Fred Saad holds the University of Montreal Endowed Chair in Prostate Cancer Research. * Corresponding author. Tel.: þ1-514-890-8000x35335; fax: þ1-514227-5103. E-mail addresses: [email protected], katharina. [email protected] (K. Boehm). http://dx.doi.org/10.1016/j.urolonc.2015.07.006 1078-1439/r 2015 Elsevier Inc. All rights reserved.
1. Introduction Obesity and overweightness are on sharp rise in western countries. In 2009 to 2010, 77% of North-American men older than 60 years were classified as overweight (body mass index [BMI]: 25–29.9 kg/m2) or obese (BMI Z 30 kg/m2) [1]. Obesity is associated with an increased risk of several cancers, such as colorectal and breast cancers [2]. Anthropometric measures of body fatness include BMI, waist circumference (WC), and waist-hip ratio (WHR).
2
K. Boehm et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–7
However, inconsistent results have been reported with respect to the association between obesity and PCa risk [3–6]. Specifically, in a large case-control study from Europe [7], increased BMI was inversely related to low-grade PCa risk and no relationship was found between BMI and high-grade PCa [7]. Similar results were also found in a prospective cohort study [8]. The same study showed results in the opposite direction when obesity and high-grade PCa were examined, similar to those of other authors [9]. Conversely, a European prospective cohort study showed no association between elevated BMI and risk of PCa, regardless of grade [10]. The lack of consensus regarding the magnitude and direction of the association between BMI and PCa risk is still debatable. In consequence, it is possible that the relationship between BMI and PCa varies from one population to another or that BMI alone is a poor indicator for obesity and other anthropometric measures alone, or in combination with BMI, might better illustrate body fat distribution. Several hypotheses have been proposed to explain an inverse relationship between BMI and PCa risk. First, it may ensue from lower PCa detection rates among men with elevated BMI, due to a dilutional effect of PSA in men with BMI elevation [11–15]. Second, it is also known that the metabolic syndrome, of which abdominal obesity is an integral part, is associated with a lower risk of PCa diagnosis [16,17]. Third, obesity may also be correlated with a low physical activity, which is possibly associated with PCa [18]. However, the latest evidence linking anthropometric measures and PCa risk was recently reviewed by the World Cancer Research Fund International organization [3]. Although no conclusion could be drawn for total or nonadvanced PCa, recent data were suggestive of a dose-response trend between each of these indicators and the risk of advanced PCa. Based on this, we examined the association between various anthropometric measures of body fatness and PCa risk in a study population constituted of men from Montreal, Canada.
2. Patients and methods 2.1. Study population We relied on data from the population-based case-control study Prostate Cancer & Environment Study (PROtEuS), which recruited subjects in Montreal, Canada, between 2005 and 2012. The study design was previously described [19–21]. In brief, the study population consisted of residents of Montreal metropolitan area, aged o76 years at diagnosis or recruitment. Cases consisted of histologically confirmed, newly diagnosed PCa and were actively ascertained through pathology departments across 7 hospitals in Montreal between 2005 and 2009, covering 480% of all cases diagnosed in the catchment area. Controls had no PCa diagnosis at the time of interview. Age-matched (5-y intervals) controls from the same residential area were randomly selected. Participation rates
were 79.4% for cases and 55.5% for controls. In analyses comparing sociodemographic characteristics of participants and nonparticipants, little differences emerged. Committees of all participating institutions approved the study protocol and all subjects provided informed consent. 2.2. Measures Subjects provided information about sociodemographic characteristics, lifestyle, and their medical history, including the frequency of PSA measurements and digital rectal examinations (DRE), physical activity (not very active, moderately active, and very active), and presence of diabetes (yes/no) by interviews. Anthropometric factors included selfreported current height and weight 2 years before the index date (i.e., diagnosis or interview), referred to as recent BMI. The interviewer measured waist (1 in above the umbilicus) and hip circumferences (maximum) in a standardized fashion. Missing values were recorded in 323 and 328 patients for WC and WHR, respectively. For cases, information about Gleason scores and PSA levels at diagnosis was extracted from patient's files and pathology reports. 2.3. Statistical analyses Our analyses focused on 3 anthropometric indicators of body fatness as potential risk factors: (1) WC, (2) WHR, and (3) recent BMI. WHR was calculated by dividing the waist by the hip circumference. For WC, a cutoff of 102 cm was used to define abdominal obesity, as recommended by the World Health Organization [22]. The cutoff for analyses on WHR was set at 1.0, according to the distribution among controls (mean ¼ 0.99 and median ¼ 1.0). Sensitivity analyses were performed, using 2 different cutoffs (0.97 and 1.02), based on tertiles among controls. BMI was calculated as recent weight divided by the square of the height. BMI was categorized according to the definition of the World Health Organization, into underweight (BMI o 18.5), normal weight (BMI: 18.5–24.9), overweight (BMI: 25.0–29.9), obese class I (BMI: 30.0– 34.9), obese class II (BMI: 35.0–39.9), and extreme obesity (BMI 4 40). For analyses, we combined the underweight and normal weight categories, as well as the obese class I, class II, and extreme obesity categories. Unconditional logistic regression was used to test associations between the 3 anthropometric measures and PCa risk. All models included age at index date, PCa family history in first-degree relatives, ancestry (European, Black, Asian, other, and unknown), physician visits per year (o1, 1–3, 43, and unknown), and number of PSA measurements within the 5 years preceding PCa diagnosis/interview (r4 PSA tests, 44 PSA tests, and unknown). Multivariate logistic regression was used to assess associations with overall PCa risk. Multinomial logistic regression was applied for analyses considering low-grade (Gleason r6) or high-grade (Gleason Z7) PCa. Additional
K. Boehm et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–7
3
analyses tested the association between WC (Z102 cm) and PCa risk, after adjustment for BMI (continuous) and among subgroups stratified according to BMI categories (BMI o 25, BMI: 25.0–29.9, and BMI Z 30). The same approach was used to test the association with WHR (41 vs. r1), adjusting for and stratifying by BMI. Finally, to assess the possibility of a potential PCa detection bias, subgroup analyses were conducted restricting to individuals with PSA testing and/or DRE within 2 years before the index date. Further analyses on BMI, WC, and WHR were adjusted for diabetes and physical activity. All statistical analyses were performed using R version 3.1.0 (R Project for Statistical Computing, www.R-project.org). A 2-sided P o 0.05 was considered to be statistically significant.
Table 1 Characteristics of cases (n ¼ 1,921) and controls (n ¼ 1,982) participating in PROtEuS, Montreal, Canada, 2005–2012
3. Results
PCa family historya, n (%) No 1,728 (87.2) Yes 198 (10) Don't know 56 (2.8)
1,411 (73.5) 448 (23.3) 62 (3.2)
o0.01
Ancestry, n (%) European African Asian Other Don't know
The study consisted of 1933 PCa cases and 1994 controls. After exclusion of 24 patients with unknown BMI, 1921 PCa cases and 1982 controls remained for analyses. Cases were slightly younger than controls (Table 1). Cases were more likely than controls to have a first-degree family history of PCa and to be of African ancestry. The median number of PSA measurements within the 5 years before the index date was similar between cases and controls. The interquartile range (IQR) for WC was 88 to 107 cm (median ¼ 98 cm). The IQR for WHR was 0.65 to 1.34 (mean ¼ 0.99, median ¼ 1.0). Regarding recent BMI, expressed in kg/m2, an IQR of 24 to 29 (mean ¼ 27.0, median ¼ 26.5) was recorded (Table 1). 3.1. Association between WC and PCa Abdominal obesity, as defined by a WC Z 102 cm, was not significantly associated with overall PCa (odds ratio [OR] ¼ 1.03 [0.89–1.19]), low-grade PCa (OR ¼ 0.86 [0.71–1.03]), or high-grade PCa (OR ¼ 1.17 [0.99–1.38]) (Table 2). When adjusted for BMI, elevated WC was associated with an increased risk of overall PCa (OR ¼ 1.23 [1.05–1.46]), with the association more pronounced for high-grade PCa (OR ¼ 1.47 [1.22–1.78]) (Table 3). In analyses stratified according to BMI categories, a positive association between WC Z102 cm and high-grade PCa was observed among overweight (OR 1.41 [1.11–1.78]) as well as obese (OR 1.77 [1.11–2.82]) men. There was no statistically significant interaction between WC and categorical BMI on PCa risk (Poverweight * WC ¼ 0.8; Pobese * WC ¼ 0.9). 3.2. Association between WHR and PCa No association was found between WHR 4 1.0 and overall, as well as grade-specific, PCa risk (Table 2). Similarly, sensitivity analyses suggested no association between WHR 4 0.97 and 41.02, and PCa (data not shown). In
Age, median (IQR) 2
Controls
Cases
n ¼ 1,982
n ¼ 1,921
65 (61–70)
P value
64 (59–69)
o0.01
Recent BMI, kg/m ; median (IQR)
26.6 (24.3–29.5)
Waist circumference, cm; median (IQR)
98.1 (88–108)
99.0 (89–107)
0.9
Hip circumference, cm; median (IQR)
99.2 (89–107)
99.2 (90–107)
0.7
Waist-hip ratio, median (range)
26.8, 26.3 (24.1–29.1)
1.0 (0.65–1.29)
1.0 (0.81–1.34)
0.006
0.5
1,680 86 71 131 14
(84.8) (4.3) (3.6) (6.6) (0.7)
1,687 124 24 74 12
(87.8) (6.5) (1.2) (3.9) (0.6)
o0.01
Family income, n (%) o20,000$ 20,000–29,999$ 30,000–49,999$ 50,000–79,999$ 80,000 and more Other
240 248 461 409 428 196
(12.1) (12.5) (23.3) (20.6) (21.6) (9.9)
220 262 446 423 425 145
(11.5) (13.6) (23.2) (22) (22.1) (7.5)
0.1
Education, n (%) Elementary High School College University Other
423 576 374 608 1
(21.3) (29.1) (18.9) (30.7) (0.1)
444 576 311 588 2
(23.1) (30) (16.2) (30.6) (0.1)
0.2
347 1,328 242 4
(18.1) (69.1) (12.6) (0.2)
0.03
Annual physician visits 5 years ago, n (%) oonce 385 (19.4) 1–3 1,389 (70.1) 43 208 (10.5) Unknown 0 (0) No of PSA 4.1, 5.0 (2.0–5.0) measurements,b mean, median (IQR) a
4.4, 5.0 (2.0–5.0) o0.01
First-degree relative with PCa. Within the last 5 years before the index date.
b
analyses of WHR 4 1.0 adjusted for BMI, a weak albeit statistically significant association for high-grade PCa risk was recorded (OR ¼ 1.20 [1.01–1.43]) (Table 4). In subgroup analyses according to BMI categories, no significant association between WHR and PCa risk was observed (Table 4). No interaction was found between categories of WHR and BMI, and PCa risk (Poverweight * WHR ¼ 0.3; Pobese * WHR ¼ 0.2).
4
Controls, n ¼ 1,982
Overall PCa
Low-grade PCab
High-grade PCac
Restricted to participants who have been screenedd
Cases, n ¼ 1,921
OR (95% CI)
Cases, n ¼ 807
OR (95% CI)
Cases, n ¼ 1,114
OR (95% CI)
Controls, n ¼ 1,505
Cases, n ¼ 1,903
OR (95% CI)
Waist circumference, cme o102 1,073 (54.1) Z102 711 (35.9)
1,074 (55.9) 722 (37.6)
1.00 (Ref.) 1.03 (0.89–1.19)
485 (60.1) 273 (33.8)
1.00 (Ref.) 0.86 (0.71–1.03)
589 (52.9) 449 (40.3)
1.00 (Ref.) 1.17 (0.99–1.38)
816 (41.2) 551 (27.8)
1,066 (55.5) 717 (37.3)
1.00 (Ref.) 1.03 (0.89–1.20)
WHRf r1.0 41.0
1,077 (54.3) 705 (35.6)
1,090 (56.7) 703 (36.6)
1.00 (Ref.) 1.03 (0.89–1.19)
474 (58.7) 284 (35.2)
1.00 (Ref.) 0.95 (0.79–1.14)
616 (55.3) 419 (37.6)
1.00 (Ref.) 1.09 (0.92–1.28)
817 (41.2) 550 (27.7)
1,084 (56.4) 696 (36.2)
1.00 (Ref.) 0.95 (0.82–1.10)
609 (30.7) 940 (47.4) 433 (21.8)
648 (33.7) 922 (48.0) 351 (18.3)
1.00 (Ref.) 0.87 (0.74–1.01) 0.72 (0.60–0.87)**
277 (34.3) 381 (47.2) 149 (18.5)
1.00 (Ref.) 0.83 (0.68–1.00) 0.71 (0.55–0.90)**
371 (33.3) 541 (48.6) 202 (18.1)
1.00 (Ref.) 0.89 (0.75–1.06) 0.73 (0.59–0.91)**
436 (22.0) 744 (37.5) 325 (16.4)
644 (33.5) 905 (47.1) 354 (18.4)
BMI, kg/m2 o25 25–29.9 Z30
1.00 (Ref.) 0.83 (0.71–0.98)* 0.71 (0.58–0.87)**
a Models were adjusted for age, ancestry, first-degree family history of prostate cancer, annual physician visits, and number of PSA tests within 5 years before index date (except subgroup restricted to screened individuals). b Low-grade prostate cancer defined as Gleason grade r6. c High-grade prostate cancer defined as Gleason grade 47. d Screening defined as PSA test and/or DRE within 2 years before index date. e Missing values for 323 patients. f Missing values for 328 patients. * P o 0.05. ** P o 0.01.
K. Boehm et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–7
Table 2 Adjusteda odds ratio (OR) and 95% CI for the association between waist circumference, waist-hip ratio, body mass index, and incident prostate cancer, PROtEuS, Montreal, Canada, 2005–2012
K. Boehm et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–7
5
Table 3 Adjusteda odds ratios (OR) and 95% CI for the association between waist circumference and incident PCa, adjusted and stratified according to body mass index, PROtEuS, Montreal, Canada, 2005–2012 Waist circumference, cm
Adjusted for BMI o102 Z102
Controls, n ¼ 1,784
1,073 (60.1) 711 (39.9)
Overall PCa
Low-grade PCab
High-grade PCac
Cases, n ¼ 1,796
OR (95% CI)
Cases, n ¼ 758
OR (95% CI)
Cases, n ¼ 1,038
OR (95% CI)
1,074 (59.8) 722 (40.2)
1.00 (Ref.) 1.23 (1.05–1.46)
485 (64.0) 273 (36.0)
1.00 (Ref.) 0.97 (0.78–1.20)
589 (56.7) 449 (43.3)
1.00 (Ref.) 1.47 (1.22–1.78)***
532 (29.6) 69 (3.8)
1.00 (Ref.) 1.24 (0.81–1.88)
227 (29.9) 30 (4.0)
1.00 (Ref.) 1.28 (0.77–2.14)
305 (29.4) 39 (3.8)
1.00 (Ref.) 1.20 (0.75–1.93)
470 (26.2) 400 (22.3)
1.00 (Ref.) 1.17 (0.95–1.43)
218 (28.8) 142 (18.7)
1.00 (Ref.) 0.89 (0.69–1.16)
252 (24.3) 253 (24.4)
1.00 (Ref.) 1.41 (1.11–1.78)**
72 (4.0) 258 (14.4)
1.00 (Ref.) 1.27 (0.88–1.83)
40 (5.3) 101 (13.3)
1.00 (Ref.) 0.87 (0.55–1.38)
32 (3.1) 157 (15.1)
1.00 (Ref.) 1.77 (1.11–2.82)*
BMI categories o25 kg/m2 (underweight or normal weight) o102 481 (27.0) Z102 52 (2.9) 25–29.9 kg/m2 (overweight) o102 491 (27.5) Z102 365 (20.5) Z30 kg/m2 (obese) o102 101 (5.7) Z102 294 (16.5)
a Models were adjusted for age, ancestry, first-degree family history of prostate cancer, annual physician visits, and number of PSA tests within 5 years before index date. b Low-grade prostate cancer defined as Gleason grade r6. c High-grade prostate cancer defined as Gleason grade Z7. * P o 0.05. ** P o 0.01. *** P o 0.001.
3.3. Association between recent BMI and PCa Compared with men underweight or of normal weight, overweight men were at lower risk of PCa (OR ¼ 0.87 [0.74–1.01]). Corresponding ORs for low-grade and highgrade PCa were 0.83 (0.68–1.00) and 0.89 (0.75–1.06), respectively (Table 2). The OR for overall PCa was 0.72 (0.60–0.87) for obese men compared to normal weighted men. Obesity was also inversely related to low-grade (OR ¼ 0.71 [0.55–0.90]) and high-grade (OR ¼ 0.73 [0.59–0.91]) PCa (Table 2). In all of the earlier analyses, further adjustments for PSA and DRE testing had minimal effect on results, nor did analyses restricted to individuals with PSA and/or DRE testing within 2 years before the index date, as well as adjustment for diabetes and physical activity.
4. Discussion The Review Panel from the Continuous Update Project of the World Cancer Research Fund International [3] recently concluded that greater body fatness is probably a cause of advanced PCa. This is an important observation as modifiable risk factors for PCa could not yet be identified. Nevertheless, additional evidence based on large, wellconducted studies is needed to be able to ascribe body fatness as a definite cause. The PROtEuS database enabled us to assess associations between several indicators of fat
mass, which can, independently or in combination, be associated with incident PCa. It also provided us with the opportunity to conduct analyses according to PCa grade, as specific relationships may exist according to PCa grade strata [23,24]. No significant association between WC, WHR, and PCa risk was observed in our data before adjustment for BMI. Similar to these, no association between WC or WHR and PCa risk (unadjusted for BMI) was found in the ProtecT cohort [7]. Once BMI was adjusted for in our data, excess risks of high-grade PCa were observed among men with elevated WC and, to a lesser extent, WHR. As the combination of WC or WHR with BMI might better predict PCa risk than when considered independently, we further tested associations for WC and WHR according to 3 frequently used BMI categories, as previously reported [10]. An elevated WC was associated with a higher risk of high-grade PCa in both overweight and obese individuals. Contrastingly, there was no clear evidence of an association between elevated WHR and PCa across BMI categories. Both WC and WHR have been used in the past as indices of central obesity. However, they do not necessarily predict health outcomes equally [25]. WC is an indicator of visceral fat mass, and is closely related to the percentage of fat mass, particularly in men [26]. It may thus be that the WC captures better than WHR the adiposity distribution that relates to PCa risk. Additionally, our results showed that as compared with men underweight or of normal weight based on BMI values, men who were overweight or obese had a lower
K. Boehm et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–7
6
Table 4 Adjusteda odds ratios (OR) and 95% CI for the association between waist-hip ratio and incident prostate cancer, adjusted and stratified according to body mass index, PROtEuS, Montreal, Canada, 2005–2012 Waist-hip ratio
Adjusted for BMI r1.0 41.0
Controls, n ¼ 1,782
1,077 (60.4) 705 (39.6)
Overall PCa
Low-grade PCab
High-grade PCac
Cases n ¼ 1,793
OR (95% CI)
Cases n ¼ 758
OR (95% CI)
Cases n ¼ 1,035
OR (95% CI)
1,090 (60.8) 703 (39.2)
1.00 (Ref.) 1.13 (0.97–1.31)
474 (62.5) 284 (37.5)
1.00 (Ref.) 1.03 (0.85-1.25)
616 (59.5) 419 (40.5)
1.00 (Ref.) 1.20 (1.01–1.43)
(26.1) (7.4)
1.00 (Ref.) 1.33 (0.97–1.84)
199 (26.3) 58 (7.7)
1.00 (Ref.) 1.40 (0.95–2.07)
269 (26.0) 74 (7.2)
1.00 (Ref.) 1.29 (0.90–1.85)
(27.4) (20.8)
1.00 (Ref.) 1.09 (0.89–1.34)
210 (27.7) 150 (19.8)
1.00 (Ref.) 1.00 (0.77–1.30)
281 (27.1) 223 (21.5)
1.00 (Ref.) 1.16 (0.92–1.46)
(7.3) (11.0)
1.00 (Ref.) 0.98 (0.72–1.35)
65 (8.5) 76 (10.0)
1.00 (Ref.) 0.76 (0.51–1.14)
66 (6.4) 122 (11.8)
1.00 (Ref.) 1.21 (0.83–1.76)
BMI categories 1) o25 kg/m2 (underweight or normal weight) r1.0 434 (24.4) 468 41.0 99 (5.6) 132 2) 25–29.9 kg/m2 (overweight) r1.0 492 (27.6) 491 41.0 362 (20.3) 373 3) Z30 kg/m2 (obese) r1.0 151 (8.4) 131 41.0 244 (13.7) 198
a Models were adjusted for age, ancestry, first-degree family history of prostate cancer, annual physician visits, and number of PSA tests within 5 years before index date. b Low-grade prostate cancer defined as Gleason grade r6. c High-grade prostate cancer defined as Gleason grade Z7.
risk of PCa, regardless of grade. Our findings concerning low-grade PCa risk are in line with the results from the Prostate Cancer Prevention Trial (PCPT), a prospective cohort study from the United States of America [8], showing an inverse association between BMI and lowgrade PCa. They are also in agreement with contemporary data provided by a large nested case-control study from the ProtecT (Prostate testing for cancer and Treatment) trial in the United Kingdom, where BMI was associated with a decreased risk of low-grade PCa [7]. The current evidence overall supports a positive relationship between markers of abdominal obesity and risk of advanced PCa [3] and our results support that abdominal obesity is associated with high-grade PCa risk. Some of the differences in part of our results vs. those previously reported may be due to population differences. For example, the proportion of high-grade PCa cases was higher in our study population (13%) than in the SELECT cohort (7%) or in the PROTECT cohort (5%). Patients of the current study were older, on average (64 y) compared with those in the European Prospective Investigation Into Cancer and Nutrition (EPIC) trial (53 y) and SELECT cohort (52 y). Nevertheless, several characteristics of our population were comparable to those in other populations. For example, the proportion of obesity in the current study (20%) was comparable to those reported in the EPIC trial (20%) and the PCPT ( 23%). The EPIC trial originated from 8 European countries and recruited subjects between 1992 and 2000. The PCPT stems from the United States and recruited subjects between 1994 and 2003. Conversely, the SELECT cohort, originating from the United States as well as Puerto Rico and Canada, is more recent (2001–2008),
and it reports a prevalence of obesity of nearly 30%. Our findings may also explain differences in the results of the aforementioned cohort studies. Body fat distribution might be different across populations, even within the same BMI category. For example, in some populations abdominal obesity might most frequently be present in individuals with a high BMI, whereas in other populations abdominal obesity might be present also in men with normal BMI. A possible explanation for this could be dietary pattern, which differs across populations. The present study has important strengths: its large sample size, including some 2000 histologically confirmed incident PCa cases, ability to control for potential confounding factors, especially PCa screening history, and valid and reliable information about anthropometric measurements and assessment of PCa risk according to cancer aggressiveness. Another major advantage of the current study is its population-based design. Our study also has some limitations. Reporting errors based on self-reports cannot be ruled out regarding height and weight, and this would have influenced BMI. However, most validity studies comparing current reported and measured values for weight and height indicate very high correlation coefficients [27,28]. Self-reported values thus appear sufficiently accurate so that their error would have minimal effect on epidemiological measures of association [29]. Nonetheless, other anthropometric measurements, such as weight and hip circumferences, were measured by interviewers following a preestablished protocol. Another limitation relates to response rates, which were relatively high for cases and less for controls. However, in analyses comparing sociodemographic characteristics of participants
K. Boehm et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–7
and nonparticipants, little differences emerged, reassuring against the presence of selection bias in the study. Finally, other unmeasured confounders could be at play in assessing obesity measurements, as well as PCa risk.
[11]
[12]
5. Conclusion Elevated BMI was associated with a lower risk of PCa, regardless of grade. Contrastingly, abdominal obesity, when adjusted for BMI, yielded results in the opposite direction, i. e., an increased risk of high-grade PCa. Taken together, our observations suggest that the specific body fat distribution (abdominal), for a given BMI, is a predictor of PCa risk, whereas BMI alone is not. Moreover, based on our data, WC appears to a better indicator of central obesity than WHR in predicting PCa risk. In consequence, future studies on this topic should consider both abdominal and overall obesity levels. From a clinical perspective, taking into account several anthropometric indicators conjointly such as BMI and WC may lead to a better prediction of PCa risk. References [1] Flegal KM, Carroll MD, Kit BK, Ogden CL. Prevalence of obesity and trends in the distribution of body mass index among US adults, 1999–2010. J Am Med Assoc 2012;307:491–7. [2] Renehan AG, Tyson M, Egger M, Heller RF, Zwahlen M. Body-mass index and incidence of cancer: a systematic review and meta-analysis of prospective observational studies. Lancet 2008;371:569–78. [3] WCRFI. Continuous Update Project Report: Diet, Nutrition, Physical Activity, and Prostate Cancer. Available at: http://www.wcrf.org/ sites/default/files/Prostate-Cancer-2014-Report.pdf 2014, [accessed 04.12.14]. [4] Baillargeon J, Platz EA, Rose DP, et al. Obesity, adipokines, and prostate cancer in a prospective population-based study. Cancer Epidemiol Biomarkers Prev 2006;15:1331–5. [5] Baillargeon J, Rose DP. Obesity, adipokines, and prostate cancer (review). Int J Oncol 2006;28:737–45. [6] Giovannucci E, Rimm EB, Liu Y, et al. Body mass index and risk of prostate cancer in U.S. health professionals. J Natl Cancer Inst 2003; 95:1240–4. [7] Dimitropoulou P, Martin RM, Turner EL, et al. Association of obesity with prostate cancer: a case-control study within the population-based PSA testing phase of the ProtecT study. Br J Cancer. 2011;104:875–81. [8] Gong Z, Neuhouser ML, Goodman PJ, et al. Obesity, diabetes, and risk of prostate cancer: results from the prostate cancer prevention trial. Cancer Epidemiol Biomarkers Prev 2006;15:1977–83. [9] Liang Y, Ketchum NS, Goodman PJ, Klein EA, Thompson IM Jr. Is there a role for body mass index in the assessment of prostate cancer risk on biopsy? J Urol 2014;192:1094–9. [10] Pischon T, Boeing H, Weikert S, et al. Body size and risk of prostate cancer in the European prospective investigation into
[13]
[14]
[15]
[16]
[17]
[18] [19]
[20]
[21]
[22]
[23] [24] [25]
[26]
[27]
[28]
[29]
7
cancer and nutrition. Cancer Epidemiol Biomarkers Prev 2008;17: 3252–61. Rundle A, Richards C, Neugut AI. Body composition, abdominal fat distribution, and prostate-specific antigen test results. Cancer Epidemiol Biomarkers Prev 2009;18:331–6. Rundle A, Neugut AI. Obesity and screening PSA levels among men undergoing an annual physical exam. Prostate 2008;68:373–80. Culp S, Porter M. The effect of obesity and lower serum prostatespecific antigen levels on prostate-cancer screening results in American men. BJU Int 2009;104:1457–61. Banez LL, Hamilton RJ, Partin AW, et al. Obesity-related plasma hemodilution and PSA concentration among men with prostate cancer. J Am Med Assoc 2007;298:2275–80. Baillargeon J, Pollock BH, Kristal AR, et al. The association of body mass index and prostate-specific antigen in a population-based study. Cancer 2005;103:1092–5. Tande AJ, Platz EA, Folsom AR. The metabolic syndrome is associated with reduced risk of prostate cancer. Am J Epidemiol 2006;164:1094–102. De Nunzio C, Aronson W, Freedland SJ, Giovannucci E, Parsons JK. The correlation between metabolic syndrome and prostatic diseases. Eur Urol 2012;61:560–70. Kruk J, Czerniak U. Physical activity and its relation to cancer risk: updating the evidence. Asian Pac J Cancer Prev 2013;14:3993–4003. Parent ME, Goldberg MS, Crouse DL, et al. Traffic-related air pollution and prostate cancer risk: a case-control study in Montreal, Canada. Occup Environ Med 2013;70:511–8. Blanc-Lapierre A, Weiss D, Parent ME. Use of oral anticoagulants and risk of prostate cancer: a population-based case-control study in Montreal, Canada. Cancer Causes Control 2014;25:1159–66. Spence AR, Rousseau MC, Parent ME. Sexual partners, sexually transmitted infections, and prostate cancer risk. Cancer Epidemiol 2014;38:700–7. Nishida C, Ko GT, Kumanyika S. Body fat distribution and noncommunicable diseases in populations: overview of the 2008 WHO Expert Consultation on Waist Circumference and Waist-Hip Ratio. Eur J Clin Nutr 2010;64:2–5. Freedland SJ, Platz EA. Obesity and prostate cancer: making sense out of apparently conflicting data. Epidemiol Rev 2007;29:88–97. Allott EH, Masko EM, Freedland SJ. Obesity and prostate cancer: weighing the evidence. Eur Urol 2013;63:800–9. Cheong KC, Ghazali SM, Hock LK, et al. The discriminative ability of waist circumference, body mass index and waist-to-hip ratio in identifying metabolic syndrome: Variations by age, sex and race. Diabetes Metab Syndr 2015;9:74–8. Bosy-Westphal A, Geisler C, Onur S, et al. Value of body fat mass vs anthropometric obesity indices in the assessment of metabolic risk factors. Int J Obes 2006;30:475–83. Stevens J, Keil JE, Waid LR, Gazes PC. Accuracy of current, 4-year, and 28-year self-reported body weight in an elderly population. Am J Epidemiol 1990;132:1156–63. Casey VA, Dwyer JT, Berkey CS, Coleman KA, Gardner J, Valadian I. Long-term memory of body weight and past weight satisfaction: a longitudinal follow-up study. Am J Clin Nutr 1991;53: 1493–8. Rimm EB, Stampfer MJ, Colditz GA, Chute CG, Litin LB, Willett WC. Validity of self-reported waist and hip circumferences in men and women. Epidemiology 1990;1:466–73.