Metabolic Syndrome and Prostate Cancer Risk in a Population-Based Case–Control Study in Montreal, Canada

Audrey Blanc-Lapierre; Andrea Spence; Pierre I. Karakiewicz; Armen Aprikian; Fred Saad; Marie-Élise Parent


BMC Public Health. 2015;15(913) 

In This Article


The study population comprised 1937 cases (including 532 high-grade PCa) and 1995 controls. For 3.1 % of cases and 3.9 % of controls interviews were conducted with a proxy, usually the spouse.

Cases were slightly younger than controls (Table 2, p < 0.01). As expected, cases were more likely than controls to have a family history of PCa (p < 0.01), to be of Sub-Saharan ancestry (p < 0.01) and to have been screened for PCa in the last two years (p < 0.01). They were less likely to be of Greater Middle East (p < 0.01) or Asian ancestry (p < 0.01). A regular PCa screening (≥ 5 tests during the previous five years) was more often reported by low-grade than by high-grade cases (59.2 % vs 50.4 %, respectively, p < 0.01). Cases and controls were similar in terms of education, fruit and vegetable consumption, smoking habits and alcohol consumption. Cases had been more physically active than controls during adulthood (p trend = 0.06). Cases had a slightly lower BMI than controls (mean of 26.8 vs 27.2 kg/m 2 , p < 0.01), but had a similar waist circumference (98.6 vs 98.5 cm). Dyslipidemia (29.7 % among cases vs 36.4 % among controls, p < 0.01), hypertension (37.9 % vs 42.3 %, p < 0.01) and T2D (10.6 % vs 17.4 %, p < 0.01) were less frequent among cases, especially when diabetes was diagnosed more than four years before the index date or treated with metformin. Statins and 5-alpha reductase inhibitors uses were similar among cases and controls, whereas aspirin use was more frequent among controls (p = 0.04).

Overall, 28.4 % of subjects (24.9 % of cases, 31.8 % of controls) ever met MetS criteria according to the NCEP-ATPIII definition (33.8 % if considering the waist-to-height ratio), 11.0 % according to the WHO definition and 24.5 % according to the IDF definition (Table 1). Most subjects with MetS as defined by NCEP-ATPIII had a history of dyslipidemia (94.1 %), hypertension (80.7 %) and/or abdominal obesity (61.0 %), and 35.0 % had a T2D. The MetS profile was different among subjects of Sub-Saharan ancestry, with a higher proportion of T2D (53.9 %) and a lower proportion of dyslipidemia (71.8 %). Among controls, screening in the last two years was more frequent in subjects with MetS than in subjects without MetS (85.9 % vs 74.4 %, p < 0.01), while a history of prostate biopsy was reported in similar proportions (9.4 % vs 8.0 %, p = 0.32). Among cases, median PSA levels did not differ according to the presence or absence of MetS at diagnosis (MetS: 6.0 ng/mL, no MetS: 5.8 ng/mL, p Wilcoxon = 0.12).

After adjustment for age, family history of PCa, ancestry, PCa screening and family income, subjects with a history of MetS (≥3 components according to the NCEP-ATPIII definition) were at significantly lower risk of PCa (OR = 0.70 [0.60–0.82]) as compared to subjects with fewer than three MetS components. The ORs did not vary significantly according to PCa aggressiveness (low-grade: OR = 0.69 [0.58–0.82], high-grade: OR = 0.75 [0.60–0.94]).

In a multivariate model including all the components of MetS together and the same controlling factors as previously, a history of abdominal obesity (OR = 1.09 [0.94–1.27]) or hypertension (OR = 0.93 [0.79–1.08]) were not associated with PCa, but subjects with a history of type 2 diabetes (OR = 0.66 [0.53–0.81]) or dyslipidemia (OR = 0.74 [0.63- 0.86]) were still at decreased risk of PCa. The negative association observed between dyslipidemia and PCa was stronger when adding statins use in the model (OR = 0.58 [0.47–0.71]). Once adjusted on metformin use, the risk associated with T2D was reduced, although no longer significantly (OR = 0.78 [0.59–1.05]).

The statistically inverse association between a history of MetS and PCa was also observed when using other definitions for MetS or abdominal obesity, with ORs ranging from 0.54 [0.44–0.68] (WHO) to 0.75 [0.64–0.88] (IDF) (Table 1). The negative association tended to be more pronounced among men younger than 40 years at MetS onset (Fig. 1) and among men diagnosed with PCa before age 65 (Table 3). Odds ratios were similar when considering a history of MetS or MetS prevalence at a given time (or age) (data not shown). Using subjects with no MetS component as the referent category did not change the results (data not shown). The risk decreased with the number of MetS components present (Fig. 2, p trend <0.01). This risk decrease was not linear, suggesting rather a synergistic interaction of MetS components under a multiplicative model.

Figure 1.

Odds ratioa for the risk of prostate cancer according to age at metabolic syndromeb onset. aAdjusted for age, family history of prostate cancer, ancestry, prostate cancer screening and family income. bAccording to the definition of the Adult Treatment Panel III from the National Cholesterol Education Program with body mass index instead of waist circumference which was only measured at interview

Figure 2.

Odds ratioa for the risk of prostate cancer according to number of metabolic syndrome components. aAdjusted for age, family history of prostate cancer, ancestry, prostate cancer screening and family income. The referent category was represented by subjects without any metabolic disorder

While BPH was positively associated with MetS and PCa, adjusting for BPH did not change the OR associated with MetS. Similar results were observed after exclusion of subjects with T2D, of subjects not screened for PCa in the last two years, or never screened with DRE in the last five years (Table 3). Odds ratios were of the same magnitude among the 217 subjects of Sub-Saharan ancestry as compared to others.