Clinical association of metabolic syndrome, C‐reactive protein and testosterone levels with clinically significant prostate cancer

Abstract Recently, the influence that metabolic syndrome (MetS), hormonal alterations and inflammation might have on prostate cancer (PCa) risk has been a subject of controversial debate. Herein, we aimed to investigate the association between MetS‐components, C‐reactive protein (CRP) and testosterone levels, and the risk of clinically significant PCa (Sig‐PCa) at the time of prostate biopsy. For that, men scheduled for transrectal ultrasound guided biopsy of the prostate were studied. Clinical, laboratory parameters and criteria for MetS characterization just before the biopsy were collected. A total of 524 patients were analysed, being 195 (37.2%) subsequently diagnosed with PCa and 240 (45.8%) meet the diagnostic criteria for MetS. Among patients with PCa, MetS‐diagnosis was present in 94 (48.2%). Remarkably, a higher risk of Sig‐PCa was associated to MetS, greater number of MetS‐components and higher CRP levels (odds‐ratio: 1.83, 1.30 and 2.00, respectively; P < 0.05). Moreover, higher circulating CRP levels were also associated with a more aggressive Gleason score in PCa patients. Altogether, our data reveal a clear association between the presence of MetS, a greater number of MetS‐components or CRP levels >2.5 mg/L with an increased Sig‐PCa diagnosis and/or with aggressive features, suggesting that MetS and/or CRP levels might influence PCa pathophysiology.

elevated triglyceride levels and decreased levels of high density lipoproteins (HDL)-cholesterol. 5 Several mechanisms have been proposed to explain the association between PCa and MetS including hormonal alterations (eg low circulating levels of testosterone), insulin resistance (eg high insulin and IGF-1 levels) and inflammation status (eg alterations in cytokines and C-reactive protein [CRP] levels, among others inflammatoryrelated molecules). 6 In this sense, we have recently uncovered the existence of a fine, germane crosstalk between the endocrine-metabolic status and the development and homoeostasis of the prostate gland, wherein key components of the insulin, IGF1 and adipokines axes, among other, could play a relevant pathophysiological role. 7,8 In addition, it has been suggested that low levels of testosterone could be linked with the presence of abdominal obesity, and this in turn, might cause an alteration in the metabolism of fatty acids promoting insulin resistance, 9 which might be associated to PCa risk 10,11 ; however, the association between circulating testosterone levels, metabolic status and PCa progression/aggressiveness remains controversial. [12][13][14][15] Furthermore, circulating levels of CRP, one of the most useful markers to assess varying degrees of inflammation in disease states such as obesity, diabetes mellitus (DM), etc., 16 have been found to be elevated in patients with different cancer types compared to healthy patients 17 ; but the putative association between CRP levels, metabolic status, testosterone and PCa remains still unknown. 17 Therefore, based on the information mentioned above, the aim of this study was to explore the associations and clinical consequences that the inflammatory status (using CRP levels), testosterone levels and MetS may have on the diagnosis and aggressiveness of PCa using a cohort of patients with and without MetS and/or PCa.

| Population
This is an observational study over an 18-month prospective cohort, in patients who underwent ultrasound guided prostate biopsy. The study was carried out within a project approved by our Hospital Research Ethics Committee, and informed consent was obtained from all participants. Specifically, blood sample was obtained in the morning (between 8:00 and 10:00 AM) after an overnight fasting and then, the prostate biopsy was implemented according to clinical practice. The inclusion criteria for this study was the indication of the biopsy by the clinician according to clinical practice. Recommendations to obtain a biopsy were the following: (a) in the case of non-previous biopsy, suspicious findings on digital rectal examination (DRE), PSA >10 ng/mL, or PSA 3-10 ng/ mL if free PSA ratio was low (usually, <25-30%), and (b) in the case of patients with previous biopsies with persistently suspicious of PCa (ie elevated PSA, suspicious DRE, etc.). On the other hand, the exclusion criteria were: (a) wait circumference or other relevant clinical data not well-reported; (b) previously known PCa diagnosis, and (c) patients with acute infectious disease (not underwent prostate biopsy at this time).

| Clinical data
Demographics information and medical histories of each patient were obtained. Specifically, information of previous diagnoses of hypertension, DM and hypercholesterolaemia was collected, as well as family history of PCa and current usage of 5α-reductase inhibitors, metformin or statins. Moreover, each patient underwent a physical examination before the biopsy was carried out, including data of body weight (kg), height (cm) and waist circumference (cm).
Specifically, the waist circumference was obtained by measuring the abdominal girth midway between the lowest rib margin and iliac crest while the patients were in a standing position.
As mentioned above, a blood sample (10 mL) was also collected after an overnight fasting period of~8 hours. Levels 18 For the diagnosis of MetS, at least three of the following criteria had to be met:

5.
Diagnosis of elevated blood pressure or being actively treated for hypertension.

| Prostate biopsy and pathologic analysis
Transrectal prostate biopsy was carried out under local anaesthesia using a standard peri-prostatic block, a transrectal ultrasound transducer, and an 18G automated needle biopsy instrument. Usual recommendations were to take 12 cores in patients undergoing the first biopsy procedure, and a minimum of 16 biopsy cores for those who had a previous biopsy. As recently reported, 19   A Student's t-test was used for analysis of the quantitative data and a chi-squared test was used for the qualitative variables. A Pearson test was used to study the correlation between the quantitative variables. A receiver operating characteristic (ROC) curve analysis was performed to determine the best CRP levels cut-off for the diagnosis of Sig-PCa. Uni-and multivariate analyses were performed by logistic regression models to evaluate the association of the variables with PCa and Sig-PCa. ROC curve analysis was also performed to determine the predictive capability of the variables together in the total cohort and, a sub-analysis was also performed in patients with PSA <10 ng/mL. The De-long test was used to compare the area under the curve (AUC) values.
A <5% level of significance was used to decide statistically significant differences to make our conclusions comparable to those of the related research. All the analyses and graphics were performed with GraphPad prism 6, MedCalc statistical software and SPSS version 17.0.

| Population description
Clinical data of 655 patients were selected to be included in this study; however, 131 patients were excluded based on the criteria mentioned above. Therefore, a total of 524 patients were finally included in the analysis. The demographic and clinical data from this cohort of patients according to the MetS status are shown in

| Relationship between metabolic syndrome and circulating testosterone and CRP levels
Circulating levels of testosterone and CRP were analysed in the whole cohort of patients according to the individual diagnostic criteria of MetS (I, II, III, IV and V; Table 2). Interestingly, testosterone levels were significantly lower in patients that individually met each criterion of MetS compared to those that did not meet these criteria.
In contrast, only patients that met the criterion I had higher CRP levels ( Table 2).  (Table 3). Specifically, we found that a greater number of MetS criteria tended to be associated with a higher risk of PCa (P = 0.07; being a higher blood pressure the only criteria significantly associated with the risk of PCa; Table 3). Interestingly, we found that the presence of MetS, a greater number of MetS criteria, and higher circulating CRP (but not testosterone) levels were significantly associated with a higher risk of Sig-PCa. Moreover, when we analysed each MetS criterion independently, we found that only criteria I (waist circumference) and V (elevated blood pressure) were associated with higher risk of PCa (although only a trend was found for Criteria I; P = 0.07; Table 3), as well as with higher risk of Sig-PCa (Table 3). However, no association was observed between criterion I or V and GS (data not shown). Furthermore, it should be mentioned that although a strong correlation between BMI and waist circumference was observed in our cohort, we did not found any association between BMI and the risk of PCa or Sig-PCa. On the basis of these results, we next analysed whether a greater number of MetS criteria or the circulating levels of CRP were associated to GS in PCa patients.

| Influence of MetS, CRP and testosterone levels in the diagnosis of PCa
Interestingly, our results revealed that only a higher circulating CRP levels, but not number of MetS, was positively correlated with a higher GS (GS = 6, GS = 7, GS >7; P < 0.05; Figure 1).
Further exploratory analyses were carried out to evaluate the association of drug intake or levels of HbA1c, with the diagnoses of both PCa and Sig-PCa. Specifically, no significant association between HbA1c levels or statin intake and the diagnoses of PCa or Sig-PCa was observed in our cohort of patients. However, the analysis of metformin intake revealed a significant association with an increased risk of Sig-PCa even when adjusting by glucose levels and

| MetS, CRP and testosterone levels as predictive factors of PCa
On the basis the previous results indicating the association between a higher risk of Sig-PCa with the diagnoses of MetS, a greater number of MetS criteria and higher circulating levels of CRP, we next implemented a multivariable analysis to determine the strength of the independent association of these factors with the risk of being diagnosed with a Sig-PCa. To that end, a ROC curve analysis was firstly performed to determine the best CRP cut-off levels for the diagnosis of Sig-PCa, which revealed that the best value was 2.5 mg/ L for CRP (AUC 0.60; P = 0.003).
It should be mentioned that, as might be expected, a significant association was observed between the risk of detecting a higher rate of Sig-PCa in our cohort of patients and an older age (ie <60 vs 60-70, or vs >70 years old), an elevated PSA levels (ie <3 vs 3-10, vs 10-20, or vs >20 ng/mL) or, an abnormal DRE (Table 4). Conversely, this risk significantly decreased in those patients who had a larger prostatic volume and a previous negative biopsy. Therefore, based on these associations, and to accurately determine whether the presence of MetS, a greater number of MetS criteria, or circulating CRP levels might be used as predictive factors of Sig-PCa independently, we adjusted these three variables by age, family history, PSA, 5α reductase inhibitors intake, DRE, prostate volume and number of biopsies (Table 5). Remarkably, we found that the three variables analysed were significant associated with a higher risk of Sig-PCa as follow (Table 5)     Remarkably, we found that the rate of Sig-PCa diagnoses was significantly higher in patients with MetS compared with patients without MetS (P = 0.03). Furthermore, our study indicated that the presence of MetS as well as a greater number of MetS criteria was significantly associated with a higher risk of Sig-PCa. In fact, multivariate analysis ROC curve analyses revealed that the presence of MetS and a greater number of MetS criteria could be used as diagnostic factors for Sig-PCa. Consistent with our study, Bhindi et al, 28 who previously investigated the criteria of MetS as quantitative variables, also observed that the greater the number of MetS criteria met, the greater the risk that patients had of harbouring a Sig-PCa. Interestingly, when we analysed each MetS criterion individually, we found that a higher waist circumference and elevated blood pressure (criteria I and V, respectively) were the only two factors significantly associated with an increased risk of PCa and of Sig-PCa in our cohort of patients, which is further supported by previous meta-analysis published on this specific topic. 40 In this sense, it should be mentioned that, although BMI has been commonly used to define obesity, BMI is probably less precise than the waist circumference which has been shown to have a stronger association with the inflammatory status and cardiovascular risk. 41 In fact, we found a strong correlation between both BMI and waist cir- Interestingly, since the use of metformin and statins and the risk of PCa is a controversial topic worldwide, [43][44][45] we also analysed this association in this study. Specifically, we did not observe an association between metformin or statins intake and the diagnosis of PCa in our cohort of patients; however, a clear association was found between metformin, but not statins, intake and the diagnose of Sig-PCa. Nevertheless, this observation should be taken with caution since, it was based on an exploratory analysis of drug intake and the presence of PCa at the time of prostate biopsy using a limited number of patients under metformin treatment and, without evaluating the period of time under the drug intake (which was not available in our cohort), being this latter parameter essential in this analysis since evidences have showed that only those patients with long-term consumption of metformin are the patients with less risk of any PCa. 44 Since the available studies focusing on the association between circulating testosterone levels and the risk of developing PCa are in many instances controversial, 46 preferable to have compiled data of multiple CRP and testosterone levels from each patient over a larger time interval rather than a sin-

CONF LICTS OF INTEREST
Nothing to declare.