Associations of adiponectin and leptin with stage and ... - Springer Link

Cancer Causes Control (2013) 24:323–334 DOI 10.1007/s10552-012-0118-4

ORIGINAL PAPER

Associations of adiponectin and leptin with stage and grade of PSA-detected prostate cancer: the ProtecT study Anya Burton • Richard M. Martin • Jeff Holly • J. Athene Lane • Jenny L. Donovan • Freddie C. Hamdy David E. Neal • Kate Tilling

•

Received: 17 July 2012 / Accepted: 26 November 2012 / Published online: 8 December 2012 Ó Springer Science+Business Media Dordrecht 2012

Abstract Purpose Obesity has been associated with an increased risk of advanced and fatal prostate cancer; adipokines may mediate this association. We examined associations of the adipokines leptin and adiponectin with the stage and grade of PSA-detected prostate cancer. Methods We conducted a nested case–control study comparing 311 men with mainly locally advanced (CT3, N1, or M1 cases) vs. 413 men with localized (T B2 & NX-0 & M0 controls) PSA-detected prostate cancer, recruited 2001–2009 from 9 UK regions to the ProtecT study. Associations of body mass index and adipokine levels with prostate cancer stage were determined by conditional

logistic regression and with grade (Gleason score C7 vs. B6) by unconditional logistic regression. Results Adiponectin was inversely associated with prostate cancer stage in overweight and obese men (OR 0.62; 95 % CI 0.42–0.90; p = 0.01), but not in normal weight men (OR 1.48; 0.77–2.82; p = 0.24) (p for interaction 0.007), or all men (OR 0.86; 0.66–1.11; p = 0.24). There was no compelling evidence of associations between leptin or leptin to adiponectin ratio and prostate cancer stage. No strong associations of adiponectin, leptin, or leptin:adiponectin ratio with grade were seen. Conclusions This study provides some evidence that adiponectin levels may be associated with prostate cancer stage, dependent on the degree of adiposity of the man. Our results are consistent with adiponectin countering the adverse effects of obesity on prostate cancer progression.

Electronic supplementary material The online version of this article (doi:10.1007/s10552-012-0118-4) contains supplementary material, which is available to authorized users. A. Burton (&) R. M. Martin J. A. Lane J. L. Donovan K. Tilling School of Social and Community Medicine, University of Bristol, Canynge Hall, 39 Whatley Road, Bristol BS8 2PS, UK e-mail: [email protected] R. M. Martin e-mail: [email protected] J. A. Lane e-mail: [email protected] J. L. Donovan e-mail: [email protected] K. Tilling e-mail: [email protected]

J. Holly School of Clinical Sciences, Southmead Hospital, University of Bristol, Learning and Research Building, Bristol BS10 5NB, UK e-mail: [email protected] F. C. Hamdy Nuffield Department of Surgical Sciences, John Radcliffe Hospital, Oxford OX3 9DU, UK e-mail: [email protected] D. E. Neal Department of Oncology, Addenbrooke’s Hospital, University of Cambridge, Box 279 (S4), Hills Road, Cambridge CB2 0QQ, UK e-mail: [email protected]

A. Burton R. M. Martin MRC Centre for Causal Analysis in Translational Epidemiology, School of Social and Community Medicine, University of Bristol, Oakfield House, Oakfield Grove, Bristol BS8 2BN, UK

123

324

Keywords Prostate cancer Adiponectin Leptin Progression Stage Body mass index

Introduction Prostate cancer causes considerable morbidity and mortality, particularly in countries such as North America, the UK, and Australia where obesity is also a substantial health burden [1, 2]. Autopsy studies indicate that universally[60 % of men will have histological evidence of prostate cancer by the age of 85 years [3]. However, the disease is often indolent and may remain so for a long time [4]; given the high histological prevalence, the lifetime risk of a diagnosis is much lower than would be expected at 16.5 % in the US [5]. The case fatality rate of screen-detected prostate cancer has been estimated at around 16 % [6], and the results from large screening trials provide evidence of over-diagnosis and over-treatment of low-risk cancers attributable to widespread PSA-based screening programs [7, 8]. Identification of accurate markers of progression is critical to recognizing those cancers likely to progress so that aggressive treatment can be reserved for those most at risk and unnecessary treatment-related morbidity can be prevented. Obesity is an established risk factor for multiple cancer sites [9]. Meta-analyses of observational studies have found body mass index (BMI) to be positively associated with risk of advanced and fatal prostate cancer and treatment failure [10, 11]. Adipokines, biologically active polypeptides synthesized and secreted by white adipose tissue, may mediate the association between obesity and prostate cancer progression. In vitro, leptin stimulates proliferation [12], growth factor expression [13], and migration [14] of androgen independent cells, and possesses angiogenic properties [15]. In contrast, adiponectin inhibits proliferation of androgen sensitive and androgeninsensitive prostate cancer cells [16], leptin and insulin-like growth factor I (IGF-I)-induced prostate cancer cell proliferation [16], and angiogenesis [17]. If circulating levels of these hormones are found to be strongly associated with prostate cancer progression, they could be useful prognostic biomarkers. The ratio of leptin to adiponectin may also be important in tumor progression due the ability of adiponectin to inhibit leptin-induced proliferation and their opposing actions [16, 18]. The findings of epidemiological studies of adipokines and stage or grade of prostate cancer have been mixed; adiponectin was inversely associated with stage, grade, or aggressive prostate cancer in some [19–22], but not all [23–26], studies. Conversely, leptin was positively associated with stage, grade, or aggressive prostate cancer in some [21, 26–29], but not all, studies [19, 25, 30–32]. However, many of these have been small and underpowered (mean number of prostate cancer cases in the studies of leptin = 91 and of adiponectin = 47

123

Cancer Causes Control (2013) 24:323–334

[19–33]). We examined associations of body mass index, adiponectin, leptin, and leptin:adiponectin ratio with stage and grade of prostate cancer at diagnosis in a study nested within a contemporary cohort of men with PSA-detected disease.

Methods Study population Men included in this study were selected from the prostate testing for cancer and treatment (ProtecT) study. The ProtecT study is a UK population-based randomized trial evaluating three treatments for localized prostate cancer (radical prostatectomy, radical radiotherapy, and active monitoring with regular PSA measurement) [34]. Around 225,000 men aged 50–69 years from 9 different regions across the UK were invited for a PSA test between 2001 and 2009. Approximately 110,000 men attended prostate check clinics at which research nurses recorded sociodemographic data, baseline clinical data, weight, and assessed trial eligibility. The risks and benefits of a PSA test and participation in the study were explained by trial nurses [34]. Those who were eligible, and who consented to participate, had blood taken and completed a detailed diet, health and lifestyle questionnaire. Men with a PSA level C3 ng/ml and \20 ng/ml had a digital rectal examination, a second PSA and a 10-core transrectal ultrasound guided prostate biopsy. Based on biopsy tissue histology, tumors were clinically staged using the TNM system, classified as localized (T1–T2 and NX, M0) or advanced (T3–T4 or N1 or M1) and assigned a Gleason score. In the current study, we included all men identified with CT3 or N1 or M1 prostate cancer (the ‘cases’—n = 311). Evidence of metastatic prostate cancer was only found in 7 (2.3 %) of the 311 cases; as spread of disease was mainly local, men with CT3 or N1 or M1 will be referred to as locally advanced. Three hundred and eleven men with BT2 and NX, M0 (localized) prostate cancer were matched to the locally advanced cases by five-year age band and region (the ‘controls’). Resources were available to measure adipokines in a further 102 men, therefore to increase the power to detect differences, a further 102 localized cases were matched to 102 of the locally advanced cases selected by random number allocation, to include a total of 413 men with localized cancer (the extra 102 controls were preferentially selected if they had BMI data). We conducted a matched case–control study, comparing locally advanced (cases) vs. localized (controls) cancers, and an unmatched case–control study comparing high Gleason score (C7, cases) vs. low Gleason score (B 6, controls).


Exposure assessment Non-fasting blood samples were collected from the men at the prostate check clinic (a mean of 9 weeks before diagnosis), allowed to clot, centrifuged within 2 h of collection, and transferred via a cool bag to the laboratory where serum was separated and stored in 100 ll aliquots at -80 °C. Samples had undergone 3–4 freeze–thaw cycles prior to analysis (adipokine levels remain stable over at least 6 cycles [35]). Adiponectin and leptin were measured by quantitative sandwich enzyme immunoassay (ELISA) (QuantakineÒ, R&D Systems, Minneapolis, USA). Samples from locally advanced stage cases and localized stage controls were stored and assayed together, with the laboratory personal blinded to the case–control status. Where results fell out of the analytical range of the assay, they were diluted appropriately and repeated. The intra-assay coefficient of variation (CoV) was 6.4 % for adiponectin and 5.2 % for leptin and the inter-assay CoVs were 13.3 and 12.6 %, respectively. Men were usually weighed by nurses at the prostate check clinic but if nurse-measured weight was unavailable (7.9 % of men), self-reported values were used. Self-reported height based on the following question was used: How tall are you? BMI was calculated as weight in kilograms divided by height in meters squared and categorized as \25, 25–\27.5, 27.5–\30, and C30 kg/m2. Units of alcohol per week were calculated from selfreported alcohol drunk in the previous 7 days, frequency of drinking over the last year and quantity normally consumed. Exercise was calculated by applying weights to the frequency of self-reported mild, moderate, and strenuous exercise per week in anticipation of the metabolic equivalent task value, (mild 9 3) ? (moderate 9 5) ? (strenuous 9 9), to create a total weekly score [36] and then divided into tertiles based on the total distribution. Selfreported smoking status was defined as ‘never,’ ‘ex,’ or ‘current.’ Leptin to adiponectin (L:A) ratio was calculated as leptin in ng/ml divided by adiponectin in lg/ml.

325

advanced compared with localized) using conditional logistic regression, grouped by age band and region stratum, adjusted for assay plate (as a proxy for storage and assay conditions), and age at blood draw. Associations of BMI and adipokine levels with prostate cancer grade (high grade compared with low grade) were explored using unconditional logistic regression, adjusted for assay plate, study center, and age at blood draw. Fractional polynomials were fitted to the regression models to investigate nonlinear relationships [37]. Some associations appeared to be complex, in particular, the association of adiponectin and L:A ratio with prostate cancer stage; therefore, adipokines were entered into models as both categorical (quartiles) and continuous variables. For categorical analysis, a post-estimation Wald test was used to derive a p for differences across groups. Sensitivity analyses were conducted in which analyses were repeated restricted to the 311 cases and the first 311 age- and region-matched controls and to men with BMI data only and with adipokines entered into the models as tertiles. A test for interaction between BMI (\25 vs. C25 kg/m2) and adipokines with prostate cancer stage and grade was carried out to assess whether models should be stratified by BMI. Among the subset of men with complete data on smoking and diabetes, we investigated whether these variables were confounding factors by comparing models with and without controlling for smoking and diabetes status. The study had 80 % power at the 5 % significance level to detect differences of 0.21 standard deviations in adipokine levels between locally advanced vs. localized cases. For BMI stratified analyses, the study had 80 % power at the 5 % significance level to detect differences of 0.45 standard deviations in adipokine levels between locally advanced vs. localized cases. All statistical analyses were performed using Stata v11.2 (StataCorp, College Station, TX, USA). Trent Multicentre Research Ethics Committee approved the study under the auspices of ProMPT (Prostate Mechanisms of Prostate cancer and Treatment) (MREC/01/4/061) and all men gave informed consent.

Statistical methods The distribution of baseline characteristics among men with localized and locally advanced prostate cancer was compared using Chi-squared tests for categorical, and t tests for continuous, variables. Skewed variables were log-transformed before analysis (alcohol, adiponectin, leptin, and L:A ratio) and normality was assessed by inspection of histograms and qnorm plots. Leptin, adiponectin, and L:A ratio were categorized into quartiles according to the distribution in all men. Relationships between variables and adipokines were explored using linear regression. We examined associations of BMI and adipokine levels with prostate cancer stage (locally

Results The baseline characteristics of participants by prostate cancer stage are summarized in Table 1. Demographic factors and anthropometric measures did not differ substantially between men with locally advanced (CT3 or N1 or M1—see ‘Study Population’) or localized (T1–T2 and NX, M0) prostate cancer. Most men with locally advanced prostate cancer had a Gleason score of C7 (77.1 %), whereas only 16.5 % of men with localized disease had a Gleason score of C7. Median serum adiponectin levels were slightly lower in men with locally advanced cancer.

123

326

Median serum leptin levels did not differ substantially by cancer stage, but men with localized cancer were more likely to have leptin levels in the middle quartiles, while men with locally advanced cancer were more likely to have either low or high leptin levels. A similar pattern was seen for L:A ratio. Log serum leptin and log (L:A ratio) were strongly positively correlated with BMI and log(adiponectin) was negatively correlated with BMI (Table 2). Current smokers had lower log(leptin) and log(L:A ratios) and higher log(adiponectin) than never smokers. The association of smoking with adipokine levels was further explored by adjusting for BMI—the associations remained but were somewhat attenuated (adiponectin and current smoking b coefficient: 0.13 (-0.02 to 0.28), leptin b: -0.13 (-0.27 to 0.01), L:A ratio b: -0.26 (-0.47 to -0.05)). Log L:A ratio was positively associated with log(leptin) and negatively associated with log(adiponectin). Log(leptin) was negatively associated with log(adiponectin). Univariable associations were also assessed separately in ‘cases’ and ‘controls’ and no substantial differences in these associations were seen. There was some evidence from exploration of fractional polynomials that associations of adipokines, in particular adiponectin and L:A ratio with stage, were nonlinear (supplementary Table 1); therefore, we present all results by quartiles, as well as per unit. BMI was not associated with risk of locally advanced stage, or high-grade disease (Table 3). However, there was evidence of an interaction between BMI and adiponectin and L:A ratio associations with stage (p for interaction = 0.006 and 0.009, respectively); therefore, all models were repeated stratified by BMI (\25 or C25 kg/m2). There was no strong evidence that log(adiponectin) was associated with stage in all men (Table 4). When stratified by BMI, there was not strong evidence that log(adiponectin) was associated with stage in normal weight men (odds ratio (OR) 1.48, 95 % confidence interval (CI) 0.77–2.82, p 0.24), but there was evidence of an inverse association between log(adiponectin) and stage in overweight and obese men (OR 0.62, 95 % CI 0.42–0.90, p 0.01). Compared with overweight and obese men in lowest quartile of adiponectin levels, those in the second, third, and fourth quartiles had 61, 52, and 55 % reduced risk of advanced stage prostate cancer, respectively (p for difference across groups 0.10). This pattern remained when adiponectin was explored in tertiles (data available from the author). There was no strong evidence that log(adiponectin) was associated with prostate cancer grade (Table 4). Log(leptin) as a continuous variable was not associated with stage but men with leptin levels in the second quartile appeared to have an approximately 50 % reduced risk of advanced disease when compared with the lowest quartile

123


(OR 0.50, 95 % CI 0.32–0.78, p value for leptin 0.009) (Table 5). Men in the third and fourth quartiles had comparable risk with those in the first quartile. This pattern appeared to be driven by an association of leptin with stage in overweight and obese men. However, there was no strong evidence of an interaction with BMI (p for interaction = 0.17) and this association was not robust to analysis by tertiles of leptin, therefore, it may have been a spurious finding. The patterns were similar for grade, but the confidence intervals were wider. As with leptin, there was some evidence that men with an L:A ratio in the second quartile of the distribution had a reduced risk of locally advanced stage prostate cancer compared with the first quartile (OR 0.57, 95 % CI 0.37–0.89, p value for L:A ratio = 0.03) (Table 6). However, when analyzed in tertiles, no evidence of a U-shaped association was found (p = 0.69).When stratified by BMI, there was weak evidence of an inverse association of log(L:A ratio) with stage in men with a BMI \25 and a positive association with stage in men with BMI C25 kg/m2. There was no strong evidence that log(L:A ratio) was associated with prostate cancer grade. Adjusting any of the models for smoking or diabetes status did not materially change the estimates (supplementary Tables 2, 3, 4 and 5). Additionally, adjusting models of associations of log(leptin) with prostate cancer stage or grade for log(adiponectin), or of log(adiponectin) with stage or grade for log(leptin), did not alter the effect estimates. Restricting any of the analyses to men with BMI data only, or the first 311 age- and region-matched controls did not materially affect the estimates.

Discussion Associations of adipokines with prostate cancer stage appear complex in this cohort of men with PSA-detected cancer. Adiponectin was inversely associated with more advanced disease in overweight and obese men, but not in normal weight men. Men with leptin and leptin:adiponectin ratio in the second quartile had the lowest risk of locally advanced cancer (50 and 43 % lower, respectively). No strong associations with grade were seen. A large number of studies have investigated associations between body mass index and prostate cancer risk, but findings have been weak and inconsistent [11]. Some large studies have sought to explore this by identifying sources of heterogeneity; Andersson et al. [38] found a weak positive association of BMI with prostate cancer incidence, but a stronger association with prostate cancer mortality in a study of 135 000 Swedish men. Analyses of the US-based Cancer Prevention Study II found increasing BMI was associated with increased risk of high-grade and metastatic


327

Table 1 Baseline characteristics by prostate cancer stage na

Locally advanced (n = 311) Age, years—mean(SD)

63.6 (4.6)

Localized (n = 413)

311

63.3 (4.7)

304

410 (99.3)

na

pb

413

0.34

63.4 (4.6)

413

0.14

707 (98.7)

0.47

48 (6.7)

All (n = 724)

Ethnicity—n (%) Caucasian

298 (98.0)

Family history—n (%) Positive

303 18 (5.9)

410 30 (7.3)

Gleason score—n (%)

310

413

B6

71 (22.9)

345 (83.5)

C7

239 (77.1)

68 (16.5)

Smoking—n (%) Never Ex Current Alcohol, units per week—median (IQR)

214

330

72 (33.6)

126 (38.2)

107 (50.0)

165 (50.0)

35 (16.4)

38 (11.8)

12.4 (6.9–22.6)

190

Exercise, weighted score tertiles—n (%)

12.8 (6.9–27.0)

110 41 (37.3)

69 (32.4)

T2(13–42)

31 (28.2)

75 (35.2)

T3 (43–500)

38 (34.5)

69 (32.4)

Median(IQR)

24 (6–56)

Height, cm—n (%)

175.9 (6.6)

211

BMI, kg/m2—n (%)

307 (42.5) 544 198 (36.4) 272 (50.0)

0.26 303

0.26

74 (13.6) 12.8 (6.9–22.6)

213

T1(0–12)

110

416 (57.5) \0.001

25 (6–47) 176.2 (6.7)

211

110 (34.1) 106 (32.8) 0.43 213

0.80

372

0.56

107 (33.1) 25 (6–50) 176.1 (6.6)

369

\25

60 (28.4)

100 (27.3)

160 (27.7)

25– \27.5

68 (32.2)

112 (30.5)

180 (31.1)

27.5– \30

36 (17.1)

87 (23.7)

C30

47 (22.3)

68 (18.5)

Mean (SD)

27.3 (3.9)

123 (21.3)

27.1 (3.7)

Adiponectin, lg/ml quartiles—n (%)

311

0.43

115 (19.9)

0.61

27.2 (3.8)

413

Q1 (0.9–4.5)

87 (28.0)

94 (22.8)

181 (25.0)

Q2 (4.5–6.5)

78 (25.1)

103 (24.9)

181 (25.0)

Q3 (6.6–9.7)

69 (22.2)

112 (27.1)

Q4 (9.7–37.2)

77 (24.8)

104 (25.2)

0.30

181 (25.0)

Mean (SD)/median(IQR)

7.5 (4.8)/6.3 (4.1–9.7)

7.6 (3.9)/7.0 (4.6–9.7)

0.24

7.5 (4.3)/6.6 (4.5–9.7)

Leptin, ng/ml quartiles—n (%)

311

181 (25.0)

413

Q1 (0.3–2.8)

88 (28.3)

93 (22.5)

181 (25.0)

Q2 (2.8–4.5)

61 (19.6)

120 (29.1)

181 (25.0)

Q3 (4.5–7.3)

77 (24.8)

104 (25.2)

Q4 (7.3–54.4)

85 (27.3)

96 (23.2)

0.02

181 (25.0)


6.0 (5.4)/4.6 (2.3–7.8)

5.9 (5.7)/4.4 (2.8–7.1)

0.68

5.9 (5.6)/4.5 (2.8–7.3)

Leptin:adiponectin ratio, quartiles—n (%)

311

181 (25.0)

413

Q1 (0.0–0.3)

86 (27.7)

95 (23.0)

181 (25.0)

Q2 (0.3–0.7)

64 (20.6)

117 (28.3)

181 (25.0)

Q3 (0.7–1.3)

74 (23.8)

107 (25.9)

Q4 (1.4–23.9)

87 (28.0)

94 (22.8)

0.05

181 (25.0)


1.3 (1.7)/0.8 (0.3–1.5)

1.1 (1.8)/0.7 (0.4–1.3)

0.73

1.2 (1.7)/0.7(0.3–1.4)

181 (25.0)

SD standard deviation, IQR interquartile range, BMI body mass index a

Total number of men with data available on each variable,

b

p values compares localized vs. advanced cases

or fatal prostate cancer but reduced risk of non-metastatic, low-grade disease [39]. These findings indicate a possible differential effect between BMI and risk of incident or

aggressive prostate cancer. This has been substantiated by meta-analyses of observational studies, revealing a stronger relationship with advanced or fatal disease (e.g., pooled

123

328


Table 2 Univariate associations between adipokines and demographic factors Log leptin Regression coefficient

Log adiponectin 95 % CI

p

Regression coefficient

Log L:A ratio

95 % CI

p

Regression coefficient

95 % CI

p

Age at PCC, per 1 year

0.01

(0.00–0.03)

0.03

0.01

(-0.00–0.02)

0.09

0.01

(-0.01–0.02)

0.51

Family historya

0.13

(-0.10–0.36)

0.26

0.08

(-0.10–0.25)

0.39

0.06

(-0.26–0.37)

0.73

Smoking (never vs. current)

0.11

(-0.02–0.25)

0.11

0.02

(-0.08–0.13)

0.67

0.09

(-0.10–0.28)

0.36

Smoking (never vs. current)

-0.26

(-0.46 to -0.06)

0.01

0.16

(0.00–0.32)

0.05

-0.42

(-0.71 to -0.14)

0.003

0.01

(-0.01–0.03)

0.23

0.00

(-0.01–0.02)

0.46

0.01

(-0.02–0.03)

0.65 0.84

Log alcohol Exercise (T1 vs. T2)

-0.02

(-0.22–0.18)

0.86

-0.05

(-0.20–0.11)

0.57

0.03

(-0.25–0.31)

Exercise (T1 vs. T3)

0.11

(-0.31-0.10)

0.31

-0.13

(-0.29–0.03)

0.10

0.03

(-0.25–0.30)

0.86

Height, per 5 cm

0.00

(-0.05–0.04)

0.93

0.01

(-0.03–0.04)

0.66

-0.01

(-0.07–0.05)

0.758

BMI, per 1 kg/m2

0.15

(0.14–0.16)

\0.001

(-0.41 to -0.22)

\0.001

-0.04

(-0.05 to -0.02)

\0.001

Log leptin Log adiponectin

-0.31

0.18

(0.17–0.20)

\0.001

1.18

(1.12–1.23)

\0.001

(-1.41 to -1.22)

\0.001

-1.31

For comparison, variables in columns were entered as the dependent and variables in rows were entered as the independent variables in the regression CI confidence interval, T tertile a

Negative family history compared with positive family history

Table 3 Associations between BMI and prostate cancer stage (locally advanced vs. localized) and grade (Gleason score [7 vs. \6) BMI (kg/m2)

OR (95 % CI) \25

25– \27.5

27.5– \30

[30

pa

Per log (unit)

0.29

1.01 (0.96–1.06)

Stage No. of advanced/localized Basic modelb Grade No. of high grade/low grade Basic model

c

60/100

68/112

36/87

47/68

1.00

1.01 (0.65–1.59)

0.68 (0.41–1.12)

1.10 (0.66–1.81)

65/95

68/111

46/77

50/65

1.00

0.95 (0.61–1.48)

0.83 (0.51–1.36)

1.08 (0.66–1.77)

211/367

229/348 0.81

1.00 (0.96–1.05)

OR odds ratio, CI confidence interval a

p for differences across groups, b conditional logistic regression, grouped by age and recruitment center stratum, adjusted for age at recruitment, c unconditional logistic regression adjusted for age at recruitment and recruitment center

relative risks (RR):1.12 for advanced vs. 0.96 for localized prostate cancer [11] and 1.15 for prostate cancer mortality [10] per 5 kg/m2 increase in BMI). No clear association of BMI with prostate cancer stage or grade was seen in this cohort although the early pre-clinical stages of disease in these men may have reduced the power to detect differences. Similarly, the range of adiposity among men was narrow (most men were overweight but few were obese). Follow-up to progression-specific outcomes such as metastases or death may reveal an association in this cohort. The pattern of association between BMI and prostate cancer progression emerging from the majority of the

123

literature has sometimes been attributed to detection bias— more obese men have lower PSA levels—which may prevent detection of early stage disease at standard thresholds for biopsy—and larger prostates—which may impair detection when a standard number of cores are taken. As a result, prostate cancers in obese men may not be detected until they are more advanced and less curable [40]. However, epidemiological and basic research has implicated adipokines, molecules synthesized, and secreted by adipocytes, as a plausible biological link between obesity and prostate cancer progression. In experimental studies, leptin has been seen to increase proliferation of DU-145 and PC-3 cells [12]; coincubation with IL-6 and IGF-I increased this proliferation


329

Table 4 Overall and BMI stratified associations between adiponectin and prostate cancer stage (locally advanced vs. localized) and grade (Gleason score [7 vs. \6) Adiponectin (lg/ml)

0.9–4.5

4.5–6.5

6.5–9.7

9.7–37.2

pa

Per log (unit)

p for interactiond

Stage Basic model No. of cases/controls

87/94

78/103

69/112

77/104

OR (95 % CI)e

1.00

0.81 (0.53–1.25)

0.67 (0.43–1.03)

0.81 (0.52–1.25)

8/17

11/20

15/30

26/33

1.00

2.10 (0.55–8.06)

1.03 (0.32–3.38)

1.77 (0.58–5.45)

311/413 0.35

0.86 (0.66–1.11)

BMI \25 No. of cases/controls OR (95 % CI) BMI C25

e

No. of cases/controls

55/67

39/73

30/69

27/58

OR (95 % CI)e

1.00

0.61 (0.34–1.08)

0.52 (0.28–0.93)

0.55 (0.30–1.02)

60/100 0.46

1.48 (0.77–2.82) 151/267

0.10

0.62 (0.42–0.90)b

0.006b

Grade Basic model No. of cases/controls

83/98

73/108

74/106

77/104

OR (95 % CI)f

1.00

0.84 (0.55–1.30)

0.81 (0.53–1.24)

0.89 (0.58–1.36)


11/14

11/20

19/26

24/35

OR (95 % CI)f

1.00

0.84 (0.27–2.61)

0.84 (0.30–2.35)

0.93 (0.35–2.49)

307/416 0.79

0.91 (0.70–1.18)

BMI \25 65/95 0.98

0.88 (0.49–1.57)

BMI C25 No. of cases/controls

51/71

40/72

39/59

34/51

OR (95 % CI)f

1.00

0.75 (0.44–1.30)

0.89 (0.51–1.55)

0.86 (0.48–1.55)

164/253 0.79

0.90 (0.63–1.28)

1.00


p for differences across groups, b p \ 0.05, c p \ 0.005, d test for interaction between BMI (\25 vs.[25) and adipokines with prostate cancer stage or grade, e conditional logistic regression, grouped by age and recruitment center stratum, adjusted for age at recruitmentand assay plate, f unconditional logistic regression adjusted for age at recruitment, assay plate and recruitment center

rate [41] and co-incubation with adiponectin inhibited the leptin-induced growth [18]. Leptin has been found to increase expression of a range of growth factors involved in proliferation, invasion, and metastasis in DU-145 and PC-3 cell lines [14], to possess angiogenic properties, both in vitro and in vivo [15] and to induce migration [14] and inhibit apoptosis in DU-145 and PC-3 cells [42]. Conversely, leptin was not found to stimulate proliferation of androgendependent LNCaP cells [18, 41], indicating possible differential effects on non-aggressive and aggressive prostate cancer phenotypes. Adiponectin has been found to inhibit proliferation of prostate cancer cells in vitro regardless of their androgen sensitivity [16, 18], and to antagonise the proliferative effects of leptin and IGF-I on androgen independent (DU-145) cells and dihydrotestosterone on androgen-dependent (LNCaP) cells [16]. Adiponectin is a potent inhibitor of angiogenesis; adiponectin inhibited fibroblast growth factor (FGF)-2-induced proliferation and FGF-2induced apoptosis in a range of endothelial cells in vitro, and inhibited FGF-2-induced neovascularization in vivo [17]. Previous observational studies have explored associations of adipokines with stage, grade, or other measures of aggressive prostate cancer but many were small and underpowered to provide robust evidence. Most larger

studies investigating leptin in prostate cancer progression—irrespective of whether leptin was measured pre- or post-diagnostically or the type of control group used— found no association with stage [31], grade [19, 25, 30], or aggressive [19, 33] disease. In a moderately sized crosssectional study (151 high volume and 48 low volume cases), Chang et al. [27] found elevated leptin to be higher in men with high volume compared with low volume prostate cancer. Smaller studies had more heterogeneous findings [21, 28, 29, 32, 43] indicating the possibility of publication bias. No studies exploring leptin as a categorical variable found the U-shaped association seen in our study. Adiponectin, measured prospectively in the Physician’s Health Study, was inversely associated with highgrade and lethal prostate cancer (compared with healthy controls). The authors also found evidence of an interaction with BMI—adiponectin was inversely associated with mortality in overweight men, but not normal weight men. Other studies investigating adiponectin in progression also compared levels in men with advanced or high-grade prostate cancer to those with localized or low-grade disease. Of these studies, Baillargeon et al. [25] did not find prospectively measured adiponectin to be associated with grade. Freedland et al. [23] also did not find adiponectin to

123

330


Table 5 Overall and BMI stratified associations between leptin and prostate cancer stage (locally advanced vs. localized) and grade (Gleason score C7 vs. B6) Leptin (ng/ml)

0.3–2.8

2.8–4.5

4.5–7.3

7.3–54.4

pa

Per log (unit)

0.009b

0.98 (0.80–1.20)

p for Interactiond


88/93

61/120

77/104

85/96

OR (95 % CI)e

1.00

0.50 (0.32–0.78)

0.83 (0.54–1.28)

0.96 (0.62–1.49)

38/54

16/30

5/12

1/4

1.00

0.84 (0.36–1.96)

0.47 (0.13–1.72)

0.39 (0.03–4.44)

311/413


e


27/30

29/83

48/80

47/74

OR (95 % CI)e

1.00

0.38 (0.19–0.79)

0.67 (0.34–1.33)

0.77 (0.38–1.52)

60/100 0.61

0.67 (0.37–1.21)

0.04b

1.02 (0.73–1.42)

0.07

0.98 (0.80–1.19)

151/267 0.17


85/96

68/112

67/114

87/94

OR (95 % CI)f

1.00

0.72 (0.47–1.10)

0.66 (0.43–1.01)

1.08 (0.71–1.66)


41/51

17/29

5/12

2/3

OR (95 % CI)f

1.00

0.76 (0.36–1.59)

0.55 (0.18–1.72)

0.74 (0.11–4.87)

307/416

BMI \25 65/95 0.71

0.72 (0.43–1.20)

0.21

1.10 (0.80–1.51)


25/32

38/73

45/83

56/65

OR (95 % CI)f

1.00

0.76 (0.39–1.50)

0.69 (0.36–1.34)

1.17 (0.60–2.26)

164/253 0.18


p for differences across groups, b p \ 0.05, c p \ 0.005, d test for interaction between BMI (\25 vs.[25) and adipokines with prostate cancer stage or grade, e conditional logistic regression, grouped by age and recruitment center stratum, adjusted for age at recruitment and assay plate, f unconditional logistic regression adjusted for age at recruitment, assay plate and recruitment center

be associated with stage or grade in all men, but did find it to be inversely associated with grade in overweight and obese men. Conversely, no interaction between BMI and adiponectin with stage of disease was found in a study by Sher et al. [20]. Our evidence, in combination with the majority of the published literature, indicates that adiponectin may be inversely associated with stage, grade, and mortality in overweight and obese men. Adiponectin is inversely associated with insulin resistance and subsequent hyperinsulinemia (increased insulin production). Obesity is a major risk factor for insulin resistance. C-peptide (a marker of insulin secretion) has been associated with risk of advanced [44] and high-grade prostate cancer [30]. Like the glucose-lowering and insulin-sensitizing drug Metformin, adiponectin activates 50 -AMP-activated protein kinase (AMPK), which inhibits mammalian target of rapamycin (mTOR) signaling [45]. Metformin use has been associated with reduced risk of prostate cancer in Caucasian men with diabetes [46], and with a survival benefit after diagnosis [47]. The insulin-lowering effect of high adiponectin levels would only be seen in men with insulin resistance, who are likely to be overweight or obese. This is a possible explanation for the inverse association of adiponectin with

123

prostate cancer only being observed in overweight and obese men. Genetic studies have found associations of polymorphisms in genes for adiponectin and AdipoR1 (the adiponectin receptor) with prostate cancer risk [48], supporting a causal effect of low adiponectin levels on prostate cancer risk. Obesity is also associated with a chronic low-grade inflammatory state which may substantially contribute to a hormonal milieu, along with other adipokines and growth factors, which favors aggressive tumor development [49]. Adiponectin is an anti-inflammatory adipokine and higher levels may oppose this adverse state. The U-shaped associations of leptin and L:A ratio were not anticipated. We could speculate that extreme values are indicative of metabolic imbalance, but it is also possible that these are spurious findings as they were not replicated by analysis of tertiles and multiple statistical tests were performed. There were several strengths to this study. The nested case control design, in which cases (locally advanced cancers) and controls (localized cancers) were selected from the same population of men undergoing asymptomatic PSA-testing and 10-core TRUS biopsy, minimizes the risk of selection bias. Variables were measured identically in cases and controls and consistently throughout the study,


331

Table 6 Overall and BMI stratified associations of leptin:adiponectin ratio with prostate cancer stage (locally advanced vs. localized) and grade (Gleason score C7 vs. B6) L:A ratio

0.0–0.3

0.3–0.7

0.7–1.3

1.4–23.9

pa

Per log (unit)

0.03b

1.04 (0.90–1.19)

p for interactiond


86/95

64/117

74/107

87/94

OR (95 % CI)e

1.00

0.57 (0.37–0.89)

0.78 (0.50–1.20)

1.04 (0.68–1.61)

38/47

15/35

5/11

2/7

1.00

0.38 (0.15–0.96)

0.45 (0.13–1.55)

0.32 (0.05–1.93)

311/413


e


23/36

32/73

41/87

55/71

OR (95 % CI)e

1.00

0.75 (0.37–1.52)

0.89 (0.45–1.76)

1.40 (0.71–2.74)

60/100 0.15

0.69 (0.45–1.04) 151/267

0.17

1.22 (0.97–1.54)

0.009b


78/103

70/110

78/103

81/100

OR (95 % CI)f

1.00

0.82 (0.53–1.25)

0.98 (0.64–1.50)

1.08 (0.71–1.65)


39/46

15/35

6/10

5/4

OR (95 % CI)f

1.00

0.48 (0.23–1.03)

0.59 (0.19–1.85)

1.39 (0.34–5.65)

307/416 0.63

1.01 (0.88–1.17)

BMI \25 65/95 0.21

0.89 (0.62–1.26)


20/39

43/61

49/79

52/74

OR (95 % CI)f

1.00

1.38 (0.70–2.75)

1.29 (0.66–2.52)

1.48 (0.76–2.89)

164/253 0.71

1.10 (0.88–1.37)

0.35


p for differences across groups, b p \ 0.05, c p \ 0.005, d test for interaction between BMI (\25 vs. C25) and adipokines with prostate cancer stage or grade, e conditional logistic regression, grouped by age and recruitment center stratum, adjusted for age at recruitment and assay plate, f unconditional logistic regression adjusted for age at recruitment, assay plate and recruitment center

and detailed measures of possible confounders were available for multivariable models. It is much larger than the majority of other studies in the area, particularly studies which have investigated adiponectin. This was a contemporary cohort comprised of PSA-detected cancers—an increasingly important fraction of men being diagnosed with prostate cancer. Adipokines were measured in blood taken from men a mean of 9 weeks before diagnosis; although the cancer would already have been present, men would have been unaware of this and therefore unlikely to have made any lifestyle changes, as a result of the cancer diagnosis, which could have affected the measured adipokine levels. There were limitations to our study: our population is predominately White British and so our findings cannot be applied to the whole population at risk. Adipokine measures were taken at only one time point and may not be representative of long-term exposure. A study that sought to investigate the stability of adiponectin and leptin found the intra-class correlation coefficient over varying amounts of time to be high [50]. However, the variation associated with a single exposure measure will tend to bias results toward the null; therefore, associations will not be over-

estimated. As with all observational studies, it is possible that true associations lie in unmeasured confounders—for example, to exclude the possibility of adiponectin acting as a proxy for hyperinsulinemia, measurement of a marker of insulin production such as C-peptide would be needed. Further to this, additional analysis of other hormones which may contribute to a hormonal milieu that promotes prostate cancer progression—such as androgens, estrogens and IGFs—would be a useful extension of this study. There is no healthy control group which could have increased differences in circulating adipokines because of a greater contrast with advanced cases and make our results more comparable with some other studies. However, the absence of cancer in controls in ProtecT has not been biopsy confirmed. Thompson et al. [51] found the prevalence of prostate cancer among men with PSA B3 ng/ml in the prostate cancer prevention trial to be significant (13.5 %), therefore inclusion of a control group that does not have biopsy confirmed absence of cancer introduces a systematic bias. There is a possibility of control selection bias, but results were robust to analysis excluding the 102 controls preferentially selected if they had BMI data and to analysis excluding men without BMI data. Prognosis is highly

123

332

variable for Gleason score 7 depending on whether the predominant histological pattern is grade 3 or 4 [52, 53]. Therefore, if this information had been available, we would ideally have further divided Gleason 7 into 3 ? 4 or 4 ? 3 for analyses. The early, pre-clinical presentation of cases is both a limitation and strength of this study. Although the focus of this study is on associations with stage, there are few metastatic cases and men were only included in ProtecT if they had PSA levels \20 ng/ml; thus, very advanced cases would have been excluded. The advanced cases in this study are several years pre-clinical presentation and may not be comparable with clinically advanced cases. The limited breadth of stages also somewhat restricts our ability to detect associations, due to reduced heterogeneity in the outcome. Follow-up when the trial outcomes are available in 2016 would provide important further evidence on the association of adipokines with prostate cancer progression. It is also strength, as tumor load is small which should minimize reverse causality (where, for example, the presence of cancer may affect appetite, body mass, physical activity, and/or adipokine levels). This study provides some evidence that adipokine levels may be associated with prostate cancer stage. As the associations were modest and often complex, this study does not indicate that adipokines would be useful biomarkers of prostate cancer progression. However, this work could provide valuable insight into the etiology of progression and inform therapeutic and prevention strategies. For example, increasing adiponectin levels and insulin sensitivity through diet, exercise, and weight control [30] may help protect against prostate cancer progression. Acknowledgments The authors would like to acknowledge the tremendous contribution of all members of the ProtecT study research group, and especially the following who were involved in this research. Research nurses (recruitment, sample collection, and follow-up): lead: Sue Bonnington, Lynne Bradshaw, Debbie Cooper, Emma Elliott, Pippa Herbert, Peter Holding, Joanne Howson, Mandy Jones, Teresa Lennon, Norma Lyons, Hilary Moody, Claire Plumb, Tricia O’Sullivan, Liz Salter, Sarah Tidball, Pauline Thompson; others: Tonia Adam, Sarah Askew, Sharon Atkinson, Tim Baynes, Jan Blaikie, Viv Breen, Sean Bryne, Jo Bythem, Jenny Clarke, Jenny Cloete, Susan Dark, Gill Davis, Rachael De La Rue, Elspeth Dewhurst, Anna Dimes, Nicola Dixon, Penny Ebbs, Ingrid Emmerson, Jill Ferguson, Ali Gadd, Lisa Geoghegan, Alison Grant, Collette Grant, Catherine Gray, Rosemary Godfrey, Louise Goodwin, Susie Hall, Liz Hart, Andrew Harvey, Chloe Hoult, Sarah Hawkins, Sharon Holling, Alastair Innes, Sue Kilner, Fiona Marshall, Louise Mellen, Andrea Moore, Sally Napier, Julie Needham, Kevin Pearse, Anna Pisa, Mark Rees, Elliw Richards, Lindsay Robson, Janet Roxburgh, Nikki Samuel, Irene Sharkey, Michael Slater, Donna Smith, Pippa Taggart, Helen Taylor, Ayesha Thomas, Nicola Trewick, Claire Ward, Christy Walker, Ayesha Williams, Colin Woodhouse, Elizabeth Wyber and others. Local investigators/clinicians: Prasad Bollina, Jim Catto, Andrew Doble, Alan Doherty, Garett Durkan, David Gillatt, Owen Hughes, Roger Kockelbergh, Howard Kynaston, Hing Leung, Edgar Paez, Alan Paul, Raj Persad, Philip Powell, Stephen Prescott, Derek

123

Cancer Causes Control (2013) 24:323–334 Rosario, Hartwig Schwaibold, David Tulloch, Mike Wallace. Pathologists: Selina Bhattarai, Neeta Deshmukh, John Dormer, John Goepel, David Griffiths, Ken Grigor, Pat Harnden, Nick Mayer, Jon Oxley, Mary Robinson, Murali Varma, Anne Warren. Research, biorepository, and data management: Leila Ayandi, Lucy Brindle, Paul Brown, Simon Collin, Michael Davis, Dan Dedman, Elizabeth Down, Ewa Dudziec, Luke Ferguson, Anne George, Vriti Hansraj, Dawn Jordan, Selena Josephs, Rajeev Kumar, Adam Lambert, Athene Lane, Thomas Ludlam, Gemma Marsden, Luke Marsden, Steven Oliver, Josh Phillips, Jane Pritchard, Laura Proctor, Peter Shiarly, Martin Taylor, Emma Turner, Eleanor Walsh, Oliver Wilkinson, Valentina Wright. Administrative support: Susan Baker, Elizabeth Bellis-Sheldon, Chantal Bougard, Joanne Bowtell, Catherine Brewer, Nicholas Christoforou, Rebecca Clark, Susan Coull, Christine Croker, Rosemary Currer, Claire Daisey, Gill Delaney, Rose Donohue, Susan Fry, Jean Haddow, Susan Halpin, Belle Harris, Barbara Hattrick, Sharon Holmes, Helen Hunt, Vicky Jackson, Mandy Le Butt, Jo Leworthy, Tanya Liddiatt, Alex Martin, Jainee Mauree, Susan Moore, Gill Moulam, Jackie Mutch, Kathleen Parker, Christopher Pawsey, Michelle Purdie, Teresa Robson, Lynne Smith, Carole Stenton, Tom (Prasad Bollina, Sue Bonnington, Debbie Cooper, Andrew Doble, Alan Doherty, Emma Elliott, David Gillatt, Pippa Herbert, Peter Holding, Joanne Howson, Mandy Jones, Roger Kockelbergh, Howard Kynaston, Teresa Lennon, Norma Lyons, Hilary Moody, Philip Powell, Stephen Prescott, Liz Salter, Pauline Thompson). Department of Health disclaimer: The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the Department of Health. Grant support: AB is recipient of an MRC 4-year PhD studentship at the MRC Centre for Causal Analysis in Translational Epidemiology. The ProtecT study was funded by the National Institute for Health Research Health Technology Assessment (NIHR HTA) program (HTA 96/20/99, ISRCTN20141297) and will be published in full in Health Technology Assessment. Additional funding for this study was provided by the University of Bristol Cancer Research Fund. Conflict of interest of interest.

The authors declare that they have no conflict

References 1. Ferlay J SH, Bray F, Forman D, Mathers C, Parkin DM (2008) GLOBOCAN 2008 v1.2, Cancer incidence and mortality worldwide: IARC CancerBase No. 10 Lyon. International Agency for Research on Cancerm, France 2. World Health Organisation (2006) Global database on body mass index. http://apps.who.int/bmi/index.jsp. Accessed 21 Aug 2012 3. Haas GP, Delongchamps N, Brawley OW, Wang CY, de la Roza G (2008) The worldwide epidemiology of prostate cancer: perspectives from autopsy studies. Can J Urol 15:3866–3871 4. Pashayan N, Duffy SW, Pharoah P et al (2009) Mean sojourn time, overdiagnosis, and reduction in advanced stage prostate cancer due to screening with PSA: implications of sojourn time on screening. Br J Cancer 100:1198–1204 5. Howlader N NA, Krapcho M, Neyman N, Aminou R, Waldron W, Altekruse SF, Kosary CL, Ruhl J, Tatalovich Z, Cho H, Mariotto A, Eisner MP, Lewis DR, Chen HS, Feuer EJ, Cronin KA, Edwards BK (eds) (2011) SEER Cancer statistics review, 1975–2008. National Cancer Institute, Bethesda, http://seer. cancer.gov/csr/1975_2008/, based on November 2010 SEER data submission

Cancer Causes Control (2013) 24:323–334 6. McGregor M, Hanley JA, Boivin JF, McLean RG (1998) Screening for prostate cancer: estimating the magnitude of overdetection. CMAJ 159:1368–1372 7. Hugosson J, Carlsson S, Aus G et al (2010) Mortality results from the Goteborg randomised population-based prostate-cancer screening trial. Lancet Oncol 11:725–732 8. Schroder FH, Hugosson J, Roobol MJ et al (2012) Prostatecancer mortality at 11 years of follow-up. N Engl J Med 366:981–990 9. Renehan AG, Tyson M, Egger M, Heller RF, Zwahlen M (2008) Body-mass index and incidence of cancer: a systematic review and meta-analysis of prospective observational studies. Lancet 371:569–578 10. Cao Y, Ma J (2011) Body mass index, prostate cancer-specific mortality, and biochemical recurrence: a systematic review and meta-analysis. Cancer Prev Res (Philadelphia, PA) 4:486–501 11. MacInnis RJ, English DR (2006) Body size and composition and prostate cancer risk: systematic review and meta-regression analysis. Cancer Causes Control 17:989–1003 12. Somasundar P, Frankenberry KA, Skinner H et al (2004) Prostate cancer cell proliferation is influenced by leptin. J Surg Res 118:71–82 13. Tokuda Y, Satoh Y, Fujiyama C, Toda S, Sugihara H, Masaki Z (2003) Prostate cancer cell growth is modulated by adipocytecancer cell interaction. BJU Int 91:716–720 14. Frankenberry KA, Somasundar P, McFadden DW, Vona-Davis LC (2004) Leptin induces cell migration and the expression of growth factors in human prostate cancer cells. Am J Surg 188:560–565 15. Sierra-Honigmann MR, Nath AK, Murakami C et al (1998) Biological action of leptin as an angiogenic factor. Science 281:1683–1686 16. Bub JD, Miyazaki T, Iwamoto Y (2006) Adiponectin as a growth inhibitor in prostate cancer cells. Biochem Biophys Res Commun 340:1158–1166 17. Brakenhielm E, Veitonmaki N, Cao R et al (2004) Adiponectininduced antiangiogenesis and antitumor activity involve caspasemediated endothelial cell apoptosis. Proc Natl Acad Sci USA 101:2476–2481 18. Mistry T, Digby JE, Desai KM, Randeva HS (2008) Leptin and adiponectin interact in the regulation of prostate cancer cell growth via modulation of p53 and bcl-2 expression. Bju Int 101:1317–1322 19. Li H, Stampfer MJ, Mucci L et al (2010) A 25-year prospective study of plasma adiponectin and leptin concentrations and prostate cancer risk and survival. Clin Chem 56:34–43 20. Sher DJ, Oh WK, Jacobus S, Regan MM, Lee GS, Mantzoros C (2008) Relationship between serum adiponectin and prostate cancer grade. Prostate 68:1592–1598 21. Arisan ED, Arisan S, Atis G, Palavan-Unsal N, Ergenekon E (2009) Serum adipocytokine levels in prostate cancer patients. Urol Int 82:203–208 22. Goktas S, Yilmaz MI, Caglar K, Sonmez A, Kilic S, Bedir S (2005) Prostate cancer and adiponectin. Urology 65:1168–1172 23. Freedland SJ, Sokoll LJ, Platz EA et al (2005) Association between serum adiponectin, and pathological stage and grade in men undergoing radical prostatectomy. J Urol 174:1266–1270 24. Housa D, Vernerova Z, Heracek J et al (2008) Adiponectin as a potential marker of prostate cancer progression: studies in organconfined and locally advanced prostate cancer. Physiol Res 57:451–458 25. Baillargeon J, Platz EA, Rose DP et al (2006) Obesity, adipokines, and prostate cancer in a prospective population-based study. Cancer Epidemiol Biomarkers Prev 15:1331–1335 26. Lopez Fontana CM, Maselli ME, Perez Elizalde RF, Di Milta Monaco NA, Uvilla Recupero AL, Lopez Laur JD (2011) Leptin

333

27. 28.

29.

30.

31.

32.

33.

34.

35.

36. 37.

38.

39.

40.

41.

42.

43.

44.

45.

increases prostate cancer aggressiveness. J Physiol Biochem 67:531–538 Chang S, Hursting SD, Contois JH et al (2001) Leptin and prostate cancer. Prostate 46:62–67 Saglam K, Aydur E, Yilmaz M, Goktas S (2003) Leptin influences cellular differentiation and progression in prostate cancer. J Urol 169:1308–1311 Basaria S, Muller DC, Carducci MA, Egan J, Dobs AS (2005) Hyperglycemia and insulin resistance in men with prostate carcinoma who receive androgen-deprivation therapy. Cancer 106:581–588 Neuhouser ML, Till C, Kristal A et al (2010) Finasteride modifies the relation between serum C-peptide and prostate cancer risk: results from the prostate cancer prevention trial. Cancer Prev Res 3:279–289 Freedland SJ, Sokoll LJ, Mangold LA et al (2005) Serum leptin and pathological findings at the time of radical prostatectomy. J Urol 173:773–776 Singh SK, Grifson JJ, Mavuduru RS, Agarwal MM, Mandal AK, Jha V (2010) Serum leptin: a marker of prostate cancer irrespective of obesity. Cancer Biomarkers 7:11–15 Stocks T, Lukanova A, Rinaldi S et al (2007) Insulin resistance is inversely related to prostate cancer: a prospective study in Northern Sweden. Int J Cancer 120:2678–2686 Lane JA, Hamdy FC, Martin RM, Turner EL, Neal DE, Donovan JL (2010) Latest results from the UK trials evaluating prostate cancer screening and treatment: the CAP and ProtecT studies. Eur J Cancer 46:3095–3101 Flower L, Ahuja RH, Humphries SE, Mohamed-Ali V (2000) Effects of sample handling on the stability of interleukin 6, tumour necrosis factor-a and leptin. Cytokine 12: 1712–1716 Godin G, Shephard RJ (1985) A simple method to assess exercise behavior in the community. Can J Appl Sport Sci 10:141–146 Royston P, Altman DG (1994) Regression using fractional polynomials of continuous covariates—parsimonious parametric modeling. Appl Stat J R St C 43:429–467 Andersson SO, Wolk A, Bergstrom R et al (1997) Body size and prostate cancer: a 20-year follow-up study among 135006 Swedish construction workers. JNCI 89:385–389 Rodriguez C, Freedland SJ, Deka A et al (2007) Body mass index, weight change, and risk of prostate cancer in the cancer prevention study II nutrition cohort. Cancer Epidemiol Biomarkers Prev 16:63–69 Freedland SJ, Platz EA (2007) Obesity and prostate cancer: making sense out of apparently conflicting data. Epidemiol Rev 29:88–97 Onuma M, Bub JD, Rummel TL, Iwamoto Y (2003) Prostate cancer cell-adipocyte interaction: leptin mediates androgenindependent prostate cancer cell proliferation through c-Jun NH2terminal kinase. J Biol Chem 278:42660–42667 Hoda MR, Popken G (2008) Mitogenic and anti-apoptotic actions of adipocyte-derived hormone leptin in prostate cancer cells. BJU Int 102:383–388 Laur JL, Elizalde RP, Crica A, et al. (2010) Relationship between obesity and adipokines with prostate cancer prevalence and aggressiveness. J Urol Conf e667 Ma J, Li H, Giovannucci E et al (2008) Prediagnostic body-mass index, plasma C-peptide concentration, and prostate cancer-specific mortality in men with prostate cancer: a long-term survival analysis. Lancet Oncol 9:1039–1047 Zakikhani M, Pollak M. (2007) Adiponectin and metformin inhibit prostate and colon cancer cell proliferation via AMP kinase (AMPK)—dependent mechanisms. In: Proceedings of the American Association for Cancer Research Annual Meeting, vol 48, p 1058

123

334 46. Wright J, Stanford J (2009) Metformin use and prostate cancer in Caucasian men: results from a population-based case-control study. Cancer Causes Control 20:1617–1622 47. Currie CJ, Poole CD, Jenkins-Jones S, Gale EAM, Johnson JA, Morgan CL (2012) Mortality after incident cancer in people with and without type 2 diabetes impact of metformin on survival. Diabetes Care 35:299–304 48. Kaklamani V, Yi NJ, Zhang K et al (2011) Polymorphisms of ADIPOQ and ADIPOR1 and prostate cancer risk. Metab, Clin Exp 60:1234–1243 49. Hsing AW, Sakoda LC, Chua SC Jr (2007) Obesity, metabolic syndrome, and prostate cancer. Am J Clin Nutr 86:843S–8857 50. Lee SA, Kallianpur A, Xiang YB et al (2007) Intra-individual variation of plasma adipokine levels and utility of single

123

Cancer Causes Control (2013) 24:323–334 measurement of these biomarkers in population-based studies. Cancer Epidemiol Biomarkers Prev 16:2464–2470 51. Thompson IM, Pauler DK, Goodman PJ et al (2004) Prevalence of prostate cancer among men with a prostate-specific antigen level\ or = 4.0 ng per milliliter. N Engl J Med 350:2239–2246 52. Stark JR, Perner S, Stampfer MJ et al (2009) Gleason score and lethal prostate cancer: does 3 ? 4 = 4 ? 3? J Clin Oncol 27:3459–3464 53. Chan TY, Partin AW, Walsh PC, Epstein JI (2000) Prognostic significance of Gleason score 3 ? 4 versus Gleason score 4 ? 3 tumor at radical prostatectomy. Urology 56:823–827