The relevance of interpreting BMI in older men increases as attention focuses accumulating research that obesity advances prostate cancer. Such information serves both to indentify biological pathways in prostate carcinogenesis and toward identifying intervention targets for outcomes research. We utilized BIA as a complementary body measure to better understand the component of BMI that most affects prostate cancer. In this large study of over 800 prostate cancer cases, we found Gleason 7 and Gleason 8-10 prostate cancers were significantly associated with both FM and FFM. Furthermore, FFM remained associated with high-grade prostate cancer after controlling for FM, and FFM appeared to mediate observed associations between BMI, WC, or FM.
Results from our investigation of BMI and WC are consistent with past studies reporting obesity increases the risk of advanced-stage or high-grade prostate cancer [4
]. Higher BMI and WC (continuous or categorized) were associated with increased risk of Gleason 7 and Gleason 8-10 prostate cancer, although a limited number of men with BMI of 35 or more may have limited observing a clear trend with Gleason 8-10 prostate cancer. WHR, in contrast, was associated with Gleason 7 cancer but not with Gleason 8-10 cancer. Furthermore, BMI, WC, and WHR were not significantly associated with Gleason 6 cancer. Interestingly, the BMI or WC associations were generally stronger for Gleason 8-10 prostate cancer than for Gleason 7 cancer, consistent with promoting a more aggressive phenotype. Obesity increases oxidative stress, estrogen and leptin levels, inflammatory responses, energy availability, and insulin insensitivity, while decreasing adiponectin [6
]. Thus, observed associations could be reasonably interpreted as deriving from the effects of excess total or centralized body adiposity.
We used BIA to investigate the relationship between body adiposity and prostate cancer. BIA is an accessible and inexpensive body composition method and provides an opportunity to confirm relationship between FM and prostate cancer. However it must be recognized that body composition estimates from BIA do not always equal those derived from DXA or other reference methods [16
]. Furthermore, the BIA prediction equations are less accurate among persons with either a very high or low BMI, and cannot address differences in hydration or electrolyte levels, or chronic renal failure [17
]. This error limits BIA as a quantitative assessment method for many individuals wishing to know their true body composition. Despite these limitations, BIA estimates of body composition are strongly correlated with those from DXA (r ≥ 0.80) [16
]. In unpublished data, we found strong correlations between foot-to-foot BIA and DXA (FM: r = 0.96, FFM: r = 0.82, %BF: r = 0.85, n = 31 men ages 40-85 years). These studies illustrate that BIA has limitations, but also that BIA is sufficient to rank-order participants, allowing us to ask if those with a higher level of exposure have a different prostate cancer risk than those with a low level.
Contrary to our original hypothesis, we found that FFM best mediated the statistical association between BMI, WC, and FM with high-grade prostate cancer. This does not mean, however, that greater FFM increases the risk of prostate cancer, but instead probably reflects shared genetic, hormonal, or nutritional factors for both the maintenance of lean body mass and prostate carcinogenesis. For example, pituitary growth hormone (GH) induces insulin-like growth factor 1 (IGF1) from the liver and muscle to increase skeletal muscle mass [20
]. GH and IGF1 levels decline with aging, concordant with loss of muscle and bone mass and an increase in fat deposition and BMI [21
], and selected studies have reported an association between prostate cancer and blood IGF1 [22
] or genetic variants in IGF1
] or IGFBP3
]. Interestingly, the PI3K/AKT/mTOR pathway plays an important role in translating IGF1 signals to protein synthesis and inhibiting muscle degradation and sarcopenia, and this pathway also regulates steroid, protein, ATP, and fatty acid synthesis critical in prostate carcinogenesis [20
]. Alternatively, androgens almost certainly play a role in prostate cancer progression [28
], and circulating testosterone levels decline with aging and are associated with centralized fat deposition and loss of lean mass [29
]. Men with sufficient androgen activity to support lean body mass and control fat deposition in aging may also have increased prostate cancer risk. Indeed, androgen and GH/IGF1 activities may both be involved, such that FFM may reflect the phenotypic marker of cumulative systemic systems involved in prostate carcinogenesis [31
]. Interestingly, this does not preclude obesity as a risk factor for high-grade prostate cancer. Many of these effects of obesity operate through the PI3K/ATK/mTOR (e.g., insulin) or LkB-AMPK/mTOR (e.g., adiponectin) pathways [26
], such that obesity affects pathways which may, in turn, support FFM retention. We also observed a non-significant protective association between FM and Gleason 6 cancer [5
] that became stronger after controlling for FFM, perhaps suggesting that FFM is somewhat obscuring the ability to identify an association between obesity and low-grade prostate cancer.
Strengths of this investigation include a study population with a sufficient number of high-grade cases for analysis. Body size and body composition measures were ascertained by trained staff and prior to diagnosis to prevent reporting error or treatment effects [10
]. Although prostate biopsy may miss a cancer, potential bias derived from latent cancer within the control group was minimized because all controls were without cancer at biopsy. BMI is associated with lower blood PSA levels and a larger prostate size, such that potential stage-specific associations between obesity and prostate cancer may be an artifact of factors that influence the ability to detect prostate cancer [32
]. However, the study design controls for selective healthcare access, and we controlled for PSA levels, prostate size, and obesity-related diseases that may affect prostate cancer detection or grading.
The study has several limitations. Although it is unlikely that FFM or FM levels are a consequence of a latent or undiagnosed prostate tumor, we cannot say with certainty whether current or past FFM and FM is most relevant. Addressing this would require a prospective study with repeated data collection during follow-up. BIA was chosen as a body composition assessment method because it is feasible for large epidemiologic studies, but as discussed above, BIA is susceptible to certain errors. Our results were not affected by the time between BIA measurement and the participant's last meal or last liquid. We were unable to evaluate the impact of kidney function, but have no reason to believe that kidney function differed between cases and controls. The majority of participants were white, and therefore our results may not generalize to other race/ethnicities. Chance findings can never be entirely ruled out. Replication is necessary to confirm our findings and to extent these results to other race/ethnic groups.