Metabolic syndrome-like components and prostate cancer risk: results from the Reduction by Dutasteride of Prostate Cancer Events (REDUCE) study

Abstract

Objective

To evaluate the relationship between number of metabolic syndrome (MetS)-like components and prostate cancer diagnosis in a group of men where nearly all biopsies were taken independent of prostate-specific antigen (PSA) level, thus minimising any confounding from how the various MetS-like components may influence PSA levels.

Subjects/Patients and Methods

We analysed data from 6426 men in the Reduction by Dutasteride of Prostate Cancer Events (REDUCE) study with at least one on-study biopsy. REDUCE compared dutasteride vs placebo on prostate cancer risk among men with an elevated PSA level and negative pre-study biopsy and included two on-study biopsies regardless of PSA level at 2 and 4 years. Available data for MetS-like components included data on diabetes, hypertension, hypercholesterolaemia, and body mass index. The association between number of these MetS-like components and prostate cancer risk and low-grade (Gleason sum <7) or high-grade (Gleason sum >7) vs no prostate cancer was evaluated using logistic regression.

Results

In all, 2171 men (34%) had one MetS-like component, 724 (11%) had two, and 163 (3%) had three or four. Men with more MetS-like components had lower PSA levels (P = 0.029). One vs no MetS-like components was protective for overall prostate cancer (P = 0.041) and low-grade prostate cancer (P = 0.010). Two (P = 0.69) or three to four (P = 0.15) MetS-like components were not significantly related to prostate cancer. While one MetS-like component was unrelated to high-grade prostate cancer (P = 0.97), two (P = 0.059) or three to four MetS-like components (P = 0.02) were associated with increased high-grade prostate cancer risk, although only the latter was significant.

Conclusion

When biopsies are largely PSA level independent, men with an initial elevated PSA level and a previous negative biopsy, and multiple MetS-like components were at an increased risk of high-grade prostate cancer, suggesting the link between MetS-like components and high-grade prostate cancer is unrelated to a lowered PSA level.

Introduction

Metabolic syndrome (MetS) is a disorder that comprises a combination of at least three of the following conditions: central obesity, elevated fasting glucose, elevated triglycerides, reduced high-density lipoprotein (HDL), and high blood pressure [1]. There are multiple definitions [2], but regardless of which is used, it is clear MetS is a global epidemic, particularly in Western society [2,3].

Prostate cancer is the second most frequently diagnosed cancer and sixth most common cause of cancer death worldwide [4]. Recent data suggests prostate cancer incidence rates are rising in most countries [5]. Although the major prostate cancer risk factors are age, race, and family history, there is growing evidence that environmental factors may also play a role. Key among these are diet and obesity [6].

Recent studies examined the association between MetS and prostate cancer risk with inconsistent conclusions. While some indicated MetS may be linked to increased prostate cancer risk [7–12], others found MetS was associated with decreased incidence [13], and yet others observed either no associations [6,14–16] or an increased risk only for high-grade prostate cancer [16,17]. These inconsistencies may partly be explained by differences in PSA screening rates between cohorts, as some MetS conditions may alter PSA levels. Specifically, both diabetes and obesity are associated with lower PSA levels [18–20]. As PSA screening is the primary tool used for prostate cancer detection, lower PSA levels due to MetS may lead to fewer biopsies, and thus MetS would appear ‘protective’ for prostate cancer.

The Reduction by Dutasteride of Prostate Cancer Events (REDUCE) study design offers a way to circumvent the above-mentioned limitation of lower PSA levels in men with MetS. REDUCE was a randomised clinical trial designed to compare the effect of dutasteride on prostate cancer incidence, among men with a negative baseline biopsy. Importantly, all subjects were required to undergo a protocol-mandated biopsy at 2 and 4 years after enrolment regardless of PSA level, thus eliminating the confounding effect of the influence of MetS on PSA levels. While data were not available for all of the classic MetS components, we used available data for four MetS-like components: obesity measured using body mass index (BMI) and self-reported history of diabetes, hypercholesterolaemia and hypertension. Although these variables are not identical to the classic MetS definitions, they represent a good approximation for central obesity, elevated fasting glucose, elevated triglycerides, and high blood pressure, respectively. Therefore, we examined whether the combined presence of MetS-like conditions was associated with increased prostate cancer risk, using the REDUCE. We hypothesised that the presence of more MetS-like components would be associated with increased prostate cancer risk.

Subjects/Patients and Methods

REDUCE was a 4-year, multicentre, double-blind, placebo-controlled study of dutasteride vs placebo for prostate cancer risk reduction. The protocol was approved by the Institutional Review Board at each site, and all participants provided written informed consent. A full description of REDUCE has previously been published [21]. Eligible participants were men aged 50–75 years who met the requirements for serum PSA level (2.5–10.0 ng/mL for ages 50–60 years; 3.0–10.0 ng/mL for ages 60–75 years) and had a single negative prostate biopsy (6–12 cores) ≤6 months before enrolment.

Patients were randomised to 0.5 mg dutasteride daily or placebo. At 2 and 4 years 10-core TRUS-guided biopsies were taken regardless of PSA level; unscheduled biopsies were taken if clinically indicated, and biopsies taken during months 19–24 and 43–48 replaced those scheduled for years 2 and 4. It should be noted that <7% of the biopsies taken were independent of the study protocol and therefore the confounding effect from these for-cause biopsies is expected to be relatively low [21]. Biopsies were read centrally and assigned Gleason sums. If a biopsy was positive, no second biopsy was performed so there are no instances where multiple positive biopsies must be considered.

Medical history of hypertension, hypercholesterolaemia, and diabetes was obtained at baseline. Height and weight were measured at baseline and used to calculate BMI (kg/m²). Race was self-reported. The serum PSA level was measured at study baseline, after enrolment but before randomisation. DRE findings and TRUS prostate volume were reported from the pre-study biopsy.

Statistical Analyses

Of the 8122 participants in REDUCE, we limited analyses to 6729 men (82.8%) who underwent at least one on-study biopsy (Fig. 1); the characteristics of these patients have been described previously [22]. We defined the number of MetS-like components using the data available in REDUCE. Specifically, data were unavailable for fasting glucose, lipids, triglycerides, blood pressure, and waist-circumference. Hence, we used obesity measured as a BMI of >30 kg/m² along with self-reported hypercholesterolaemia, diabetes, and hypertension to define four MetS-like components. Patients were assigned to groups based upon the number of MetS-like components. We excluded men with missing data on the MetS-like components (206), TRUS (77), PSA level (14), or DRE (6), resulting in a study population of 6426.

Fig. 1.

Study cohort from REDUCE.

We tested the association between the number of MetS-like components (zero, one, two, or three to four) and demographic and clinical variables using chi-squared for categorical variables, anova for normally distributed continuous variables, and Kruskal–Wallis for non-normally distributed continuous variables. Variables included were chosen a priori relative to this post hoc analysis from REDUCE and included variables thought to be important predictors of prostate cancer and included age, race, geographic location, baseline PSA level, pre-study TRUS prostate volume, DRE findings, and REDUCE study arm. We also examined the frequencies of number of MetS-like components within the no prostate cancer, low-grade (Gleason sum <7), and high-grade (Gleason sum ≥7) prostate cancer groups and tested the association using chi-squared.

Logistic regression was used to examine the association between the number of MetS-like components and the risk of prostate cancer vs no prostate cancer. Multinomial regression was used to examine the association between the number of MetS-like components and the risk of low-grade prostate cancer vs no prostate cancer or high-grade prostate cancer vs no prostate cancer. We also examined the number of MetS-like components as a continuous variable in both the logistic and multinomial logistic models to test for a trend. All models were adjusted for age (continuous), race (White, Black, Other), geographic location (Europe; USA, Canada, and Puerto Rico; Other), PSA level (continuous), prostate volume (continuous; log-transformed), DRE findings (normal vs suspicious), and REDUCE study arm (dutasteride vs placebo). PSA levels in the restricted range of 2.5–10.0 ng/mL, as required by study entry criteria, were normally distributed.

To examine if the relationship between MetS-like components and prostate cancer varied by REDUCE study arm, we stratified by placebo vs dutasteride use and repeated the analyses. We also tested for an interaction between dutasteride use and the number of MetS-like components in the adjusted logistic and multinomial logistic models.

All P-values reported are two-sided and an α-level of 0.05 was considered statistically significant for all analyses. All statistical analyses were performed using Stata 11.2 (Stata Corp., College Station, TX, USA).

Results

Table 1 shows the demographic and clinical features of the 6426 participants by number of MetS-like components. Overall, there were 2171 men (34%) with one MetS-like component, 724 (11%) with two, and 163 men (3%) with either three or four. Men with more MetS-like components were more likely to have lower PSA levels (P = 0.029) and larger prostate volume (P < 0.001). Although there was a significant association between the number of MetS-like components and older age (P = 0.001), the magnitude of the differences was small. Overall race and geographic region were related to the number of MetS-like components (P < 0.001 and P = 0.012, respectively). Specifically, men from Canada, the USA, and Puerto Rico as well as Black men were more likely to have more MetS-like components compared with European men and other ethnicities.

Table 1.

Baseline demographic and clinical characteristics for patients with a given number of metabolic syndrome components.

Variable	Number of MetS components				P
	0	1	2	3–4
No. patients, N (%)	3368 (52)	2171 (34)	724 (11)	163 (3)	–
Mean (sd) age, years	62.5 (6.0)	63.1 (6.0)	62.9 (6.1)	62.7 (5.9)	0.001*
Race, N (%)					0.012†
White	3083 (92)	1989 (92)	672 (93)	149 (92)
Black	48 (1)	48 (2)	16 (2)	7 (4)
Other	237 (7)	134 (6)	36 (5)	7 (4)
Geographic region, N (%)					<0.001†
Europe	2077 (62)	1206 (55)	359 (50)	86 (53)
Canada, USA and Puerto Rico	650 (19)	599 (28)	247 (34)	58 (35)
Other	641 (19)	366 (17)	118 (16)	19 (12)
Median (IQR):					0.029‡
PSA level, ng/mL	5.8 (4.4,7.4)	5.6 (4.4,7.2)	5.7 (4.3,7.4)	5.4 (4.3,6.5)
Prostate volume, mL	42.8 (32.6,55.4)	43.9 (33.8,56.8)	46.1 (33.6,59.4)	45.9 (34.3,59.8)	<0.001‡
DRE, N (%)					0.056†
Normal	3254 (97)	2090 (96)	684 (94)	157 (96)
Suspicious	114 (3)	81 (4)	40 (6)	6 (4)
REDUCE study arm, N (%)					0.998†
Placebo	1713 (51)	1107 (51)	366 (51)	83 (51)
Dutasteride	1655 (49)	1064 (49)	358 (49)	80 (49)
Hypertension, N (%)	0	899 (41)	537 (74)	151 (93)	<0.001†
Hypercholesterolaemia, N (%)	0	448 (20)	299 (41)	106 (65)	<0.001†
Diabetes, N (%)	0	117 (5)	154 (21)	104 (64)	<0.001†
Obesity (BMI ≥30 kg/m2), N (%)	0	707 (33)	458 (63)	138 (85)	<0.001†

IQR, interquartile range. P-values calculated using

^*ANOVA,

^†chi-square, or

^‡Kruskal-Wallis.

We analysed the association between the number of MetS-like components and prostate cancer grade using chi-squared (Table 2). There was a statistically significant positive association between the number of MetS-like components and prostate cancer grade (P = 0.042) with men having two to four MetS-like components being more likely to have high-grade prostate cancer.

Table 2.

Number of MetS components in those with no prostate cancer, low-grade (Gleason sum ≤6), and high-grade prostate cancer (Gleason sum 7–10).

	No prostate cancer, N (%)	Low-grade prostate cancer, N (%)	High-grade prostate cancer, N (%)	P
Number of MetS components:				0.042
0	2587 (52)	564 (56)	217 (49)
1	1717 (34)	307 (30)	147 (33)
2	556 (11)	109 (11)	59 (13)
3–4	119 (2)	27 (3)	17 (4)

P-value calculated using chi-square.

Crude and adjusted results from the logistic regression models were similar, so only adjusted models are shown in Table 3. On multivariable analyses, men with one MetS-like component were at lower risk of overall prostate cancer (odds ratio [OR] 0.87, P = 0.041) and low-grade prostate cancer (OR 0.82, P = 0.010), compared with men with no MetS-like components. The presence of two (P = 0.510) or three to four (P = 0.706) components was not associated with risk of overall or low-grade prostate cancer vs no cancer. However, having three or four MetS-like components was significantly associated with increased risk of high-grade prostate cancer vs no cancer (OR 1.94, P = 0.017). There was a non-statistically significant trend for men with two components to be at an increased risk for high-grade prostate cancer (OR 1.35, P = 0.059). However, the presence of one MetS-like component was not associated with risk of high-grade prostate cancer (P = 0.965). When the number of MetS-like components was analysed on a continuum, there was an association between the presence of more MetS-like components and increased risk of high-grade (OR 1.16, P = 0.017), but not low-grade prostate cancer (P = 0.25).

Table 3.

Adjusted^* ORs and 95% CIs for risk of prostate cancer vs no prostate cancer, low-grade prostate cancer vs no prostate cancer and high-grade prostate cancer vs no prostate cancer in the entire cohort and by REDUCE study arm.

Study arm	Number of MetS components	All prostate cancers			Low-grade prostate cancer			High-grade prostate cancer
		OR†	95% CI	P	OR†	95% CI	P	OR†	95% CI	P
All patients	0	1.00	Ref.	–	1.00	Ref.	–	1.00	Ref.	–
	1	0.87	0.76–0.99	0.041	0.82	0.70–0.95	0.010	1.00	0.80–1.26	0.965
	2	1.04	0.86–1.26	0.690	0.93	0.74–1.16	0.510	1.35	0.99–1.84	0.059
	3–4	1.31	0.91–1.88	0.146	1.09	0.71–1.67	0.706	1.94	1.13–3.33	0.017
	Continuous	1.01	0.94–1.10	0.784	0.95	0.87–1.03	0.251	1.16	1.03–1.32	0.017
Placebo	0	1.00	Ref.	–	1.00	Ref.	–	1.00	Ref.	–
	1	0.81	0.68–0.97	0.023	0.79	0.65–0.97	0.025	0.86	0.63–1.18	0.362
	2	1.02	0.79–1.33	0.861	0.95	0.70–1.28	0.731	1.24	0.80–1.91	0.332
	3–4	1.22	0.75–2.01	0.425	1.01	0.56–1.81	0.968	1.91	0.91–4.01	0.089
	Continuous	0.99	0.89–1.09	0.785	0.94	0.84–1.06	0.321	1.11	0.93–1.32	0.249
Dutasteride	0	1.00	Ref.	–	1.00	Ref.	–	1.00	Ref.	–
	1	0.95	0.78–1.16	0.632	0.86	0.68–1.08	0.200	1.17	0.85–1.61	0.340
	2	1.06	0.80–1.42	0.674	0.89	0.63–1.27	0.532	1.48	0.95–2.31	0.085
	3–4	1.40	0.82–2.38	0.215	1.17	0.61–2.22	0.639	2.00	0.91–4.40	0.087
	Continuous	1.04	0.93–1.17	0.473	0.96	0.83–1.10	0.530	1.22	1.02–1.46	0.026

Referents (Ref.) are no prostate cancer.

^*Adjusted for age, race, geographical region, PSA level, TRUS prostate volume, DRE findings, and REDUCE study arm (in model of all patients). Significant P-values shown in bold font. All P-interactions >0.20.

When stratified by treatment arm, overall associations between MetS-like components and prostate cancer risk were similar in both arms (P-interactions >0.20; Table 3). Specifically, for both men on dutasteride and placebo, the presence of three to four MetS-like components was associated with high-grade prostate cancer, but not low-grade prostate cancer, although this only reached statistical significance in the dutasteride arm (P = 0.026). Given concerns over differential screening practices in different locations, we also examined the results stratified by geographic region and found similar results in all geographic regions (P-interactions >0.19, data not shown).

Discussion

Our present findings from the REDUCE study, wherein men with an elevated PSA level and a negative pre-study biopsy were required to undergo biopsy regardless of PSA level, revealed that having three or more MetS-like components was associated with an increased risk of high-grade prostate cancer, but not with overall or low-grade prostate cancer. Indeed, the presence of one MetS-like component was associated with a reduced risk of overall and low-grade prostate cancer. These findings provide further support for the hypothesis that metabolic abnormalities related to obesity, diabetes, hypertension, and hypercholesterolaemia are associated with aggressive prostate cancer. As such, future studies should be directed at determining if efforts to prevent MetS may have benefits in terms of reducing the risk of aggressive prostate cancer.

Overall, the present study did not find an association between MetS-like components and prostate cancer diagnosis among men with an elevated PSA level and a negative pre-study biopsy. While our definition of MetS-like components differs from many prior studies, our present results are nonetheless consistent with reports from some studies that used classic MetS definitions [15–17,23], while disagreeing with findings from others [8,9,11–13]. Our present finding agrees with those of a prospective study of 29 364 Norwegian men, followed-up to 10 years, which found a null association between MetS and incident prostate cancer diagnosis [15]. Similarly, a prospective study of 2408 men who underwent prostate biopsies in Spain, found no association between MetS and prostate cancer diagnosis [17]. Finally, a meta-analysis of 19 studies confirmed a lack of significant association between MetS and overall prostate cancer diagnosis [16].

Contrary to our present findings, a USA cohort study of 7082 men found an inverse association between MetS, number of MetS components, and overall prostate cancer risk [13]. This raises the hypothesis that the association between MetS and prostate cancer risk may vary geographically. For example, PSA screening is quite common in the USA, but much less common in Europe (although rates are increasing). Thus, if MetS influences PSA levels without altering prostate cancer risk per se, due to wide-spread PSA screening in the USA, it would appear that MetS influences prostate cancer risk. Indeed, both obesity and diabetes, two MetS components, have been linked with lower PSA levels [18–20]. As such, in the USA-based study, it is possible that MetS appeared ‘protective’ for overall prostate cancer risk due to artificially lowered PSA levels. Our study approach avoided this bias by using the REDUCE study, where men underwent a protocol-mandated biopsy at 2 and 4 years regardless of PSA level. Therefore it is possible that the inverse associations found in prior studies may have been driven by PSA level differences or screening practices.

Interestingly, a prospective study among 1880 Finnish men with an average of 13 years follow-up found a positive association between MetS and prostate cancer [9]. However, they did not stratify by Gleason sum or the number of MetS components, and diabetic patients were excluded [9]. Similarly, a prospective study of 16 209 men found two or more MetS components were associated with increased overall prostate cancer risk, but did not stratify by grade [8]. The duration of that study was greater than the REDUCE study, spanning 27 years, perhaps indicating longer follow-up may be necessary to detect significant differences in overall prostate cancer risk relative to multiple MetS components [8]. Notably, both studies used prostate cancer diagnosis from cancer registries and did not mandate biopsies or PSA screening [8,9]. Because these studies included patients not regularly screened for prostate cancer, it remains plausible the associations observed were driven by increased risk of clinically detected high-grade prostate cancer. This raises the hypothesis the association between MetS and prostate cancer risk may vary by grade.

Hence, we tested whether the association between the number of MetS-like components and prostate cancer risk varies by grade. We found the presence of more MetS-like components was associated with increased risk of high-grade, but not low-grade prostate cancer. While most prior studies regarding prostate cancer grade investigated MetS as a whole rather than the number of MetS components, they found similar results [7,12,17,23]. Importantly, we found men without MetS (i.e. three or more components), but with at least two MetS-like components may be at increased risk of high-grade prostate cancer. This suggests efforts to reduce high-grade prostate cancer should not just focus on men with MetS but include men with metabolic abnormalities before they fully develop MetS.

There are several biologically plausible mechanisms explaining why MetS may increase high-grade prostate cancer risk. MetS conditions are associated with a pro-inflammatory state (elevated levels of C-reactive protein [CRP], TNF-α, interleukin 8 [IL-8], IL-6, and IL-1β), which has been linked to prostate cancer risk [24–26]. Also, high cholesterol levels associated with MetS have been linked with increased risk of high-grade prostate cancer [27]. Interestingly, type II diabetes, a MetS component, when examined alone is associated with reduced prostate cancer risk, possibly due to a hypoinsulinaemic state, which results from damage to pancreatic β-cells [28–30]. However, hyperinsulinaemia, a common condition among men with the MetS, has been associated with increased risk of prostate cancer death [31]. Finally, MetS conditions can also alter circulating levels of IGF-1 [32], leptin [33] and adiponectin [34], all of which have been associated with prostate cancer risk [32–34]. Thus, it is clear the associations between MetS and prostate cancer risk are complex and further larger studies in multi-ethnic cohorts are warranted.

Our present findings have several important clinical implications. Because the presence of more MetS-like components is associated with aggressive prostate cancer, it is of great importance that patients maintain a healthy lifestyle. However, it is unknown whether lifestyle improvements or medications can mitigate the impact of MetS on a pre-existing high-grade prostate cancer. In another REDUCE-based study, statin medications for hypercholesterolaemia, a MetS component, were not associated with overall or high-grade prostate cancer risk [22], whereas others found statins may slow prostate cancer progression [35].

The present study has some limitations. Due to limited available data collected during the REDUCE study, we used diabetes, hypertension, hypercholesterolaemia, and BMI to approximate MetS-like components. Because our present study used diabetes instead of fasting glucose, pre-diabetic patients were excluded. As recent studies have shown that patients have increased prostate cancer risk in the early stages of diabetes and have reduced prostate cancer risk with longer duration of diabetes, this omission could bias the results to understate prostate cancer risk [36]. Information about MetS-Iike components was collected as ‘yes/no’ and all fields were self-reported with the exception of obesity, which was determined from BMI. Moreover, data on severity or duration of the MetS-like components was unknown. Thus, it is possible men with well-controlled MetS may have different risks than poorly controlled MetS. The REDUCE cohort was predominantly White and excluded patients with positive baseline biopsies. Thus, it is possible that MetS-like components are linked with larger tumours that are detected on the initial biopsy, as shown by others [12]. If true, our present findings linking the MetS-like components and high-grade prostate cancer may have underestimated the true association between metabolic abnormality and aggressive prostate cancer. Moreover, the limited number of men with Gleason sum 8–10 and with three to four MetS-like components prevented us from examining MetS-like components and very high-grade prostate cancer. Additionally, while enrolment was limited to PSA levels of 2.5–10 ng/mL, a key strength was that all participants were required to undergo biopsy. These limitations notwithstanding, our present study was equipped to investigate associations between multiple MetS-like components and overall, low-, and high-grade prostate cancer risk, in a population with a previous negative biopsy and that was biopsied regardless of PSA level.

Prevalence of MetS-like components in our present cohort was substantially lower than the published reported prevalence [3]. This is probably due to two main issues. Firstly, we had no data for HDL or triglycerides and therefore, used hypercholesterolaemia as a proxy. Consequently, we only examined four MetS-like components rather than five in classic MetS. Patients with only two MetS-like components in our present study may well have three (or more) classic MetS components. Secondly, as variables were self-reported, it is likely that a man taking a medication to control a particular component may feel that he no longer has that component when self-reporting. For example, in a prior study we found that many men who were taking a statin did not report hypercholesterolaemia [37]. This creates a misclassification in that some men with the MetS-like component of interest were considered not to have it. Thus, we considered using medications as a means to classify subjects into MetS-like components. However, as we do not know the indication for each medication (i.e. statins can also be given for high high-sensitivity CRP levels; some anti-hypertensive medications can be given for urinary complaints), this would create a problem with misclassification in the other direction (i.e. considering men as having a MetS-like component when they do not). As such, we did not include this analysis, but acknowledge that we may have under classified men. Importantly, however, as BMI, hypertension, and diabetes are all included in the definition of MetS according to the WHO, with the exception of hypercholesterolaemia we did not over classify men (i.e. assign them a MetS-like component when they did not have it). As misclassification in general tends to bias the results to the null, our observation that three to four MetS-like components is linked with high-grade prostate cancer, may actually underestimate the true association. Finally, as our analysis was post hoc relative to the REDUCE primary study analyses; it is possible our present results are due to type I error. Also, in the placebo arm, more MetS-like components were not significantly associated with high-grade prostate cancer risk and thus it is possible that the results were driven by the dutasteride arm. However, we think this is unlikely as the P-value for three to four MetS-like components in the placebo arm approached significance (P = 0.089) with a magnitude (OR 1.91) almost identical to that for the dutasteride group (OR 2.00, P = 0.087).

In summary, among men with an elevated PSA level and a negative pre-study biopsy wherein all men were prescribed to undergo biopsy independent of PSA level, men with three to four MetS-like components had an increased risk of high-grade prostate cancer. While MetS is more commonly viewed as a USA phenomenon, 44% of the REDUCE subjects in Europe had at least one MetS-like component, highlighting the global nature of the problem. Together with other data confirming the association between MetS and aggressive prostate cancer, these findings suggest metabolic abnormalities may also be a risk factor for aggressive prostate cancer. As such, efforts to reduce MetS and thereby aggressive prostate cancer are warranted.

Abbreviations

BMI

body mass index

CRP

C-reactive protein

HDL

high-density lipoprotein

IL

interleukin

MetS

metabolic syndrome

OR

odds ratio

REDUCE

Reduction by Dutasteride of Prostate Cancer Events (Study)